The Taub Faculty of Computer Science Events and Talks

...
[ Full version ]

Additive Fast Fourier Transforms and the finite fields in which they work have a main role in many algebraic applications. One of these applications is to implement a whole PCP system to obtain a succinct proof of computational integrity. In this work we modifty a known algorithm to calculate additive FFTs over affine subspaces in GF(2^k); where previous works were applied only to linear subspaces. We present a parallel implementation of this algorithm for CPU and GPU architectu...
[ Full version ]

Google will give a technical talk on Validating Gmail and Maintaining Gmail's data, on Wednesday, December 30th, 2015, at 12:30, in Class 9, Taub Building for Computer Science. Please register ahead....
[ Full version ]

The class P attempts to capture the efficiently solvable computational tasks. It is full of practically relevant problems, with varied and fascinating combinatorial structure. In this talk, I will give an overview of a rapidly growing body of work that seeks a better understanding of the structure within P. Inspired by NP-hardness, the main tool in this approach are combinatorial reductions. Combining these reductions with a small set of plausible conjectures, we obtai...
[ Full version ]

Coping with network failures has been recognized as an issue of major importance in terms of security, stability and prosperity. It has become clear that current networking standards fall short of coping with the complex challenge of surviving failures. The need to address this challenge has become a focal point of networking research. Accordingly, the goal of this research is to establish a comprehensive methodology for efficiently providing network survivability in conjunction w...
[ Full version ]

...
[ Full version ]

Collecting large amounts of data is now common in multiple applications. In certain cases, particularly involving real-time processing, we might, however, not even be able to analyze all of it. This raises the following question: How well can we perform some task under severe computational constraints. In this talk I'll discuss such challenges for two specific problems: 1) edge detection from large and noisy images; and 2) detection of strong correlation...
[ Full version ]

Defining an optimal protection strategy against viruses, spam propagation or any other kind of contamination process is an important feature for designing new networks and architectures. In this work, we consider decentralized optimal protection strategies when a virus is propagating over a network through a SIS epidemic process. We assume that each node in the network can fully protect itself from infection at a constant cost, or the node can use recovery software, once it is inf...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

Recent years have seen a surge of activity in the field of arithmetic circuit lower bounds. Although there has been a tremendous improvement in our understanding of arithmetic circuits, they do not translate to analogues in the boolean world. In this talk, In this talk, we shall look at a possible approach towards strengthening arithmetic circuit lower bounds so that they may have relevance in the boolean world, namely via `functional' lower bounds. We say that an ...
[ Full version ]

The study of security in decentralized large-scale distributed systems has made a large leap forward with the introduction of cyber-currencies, starting with Bitcoin. This talk will discuss three results tackling central challenges that were exposed by this progress. Cyber-currencies, as the name implies, implement secure and reliable digital currencies, but their promise has grown beyond currency, for example to so-called smart contracts and to digital asset inf...
[ Full version ]

Amazon Web Services (AWS) first AWSome Day will take place at the Technion on Tuesday, December 23st. between 9:00 to 16:00 at the EE Meyer Building 1003. The free one-day training event will be delivered by AWS Educational Instructors, and is ideal for technical leaders, IT Pros and Developers who are eager to learn more about getting started with the AWS Cloud. ...
[ Full version ]

It has long been conjectured that hypothesis spaces suitable for data that is compositional in nature, such as text or images, may be more efficiently represented with deep hierarchical architectures than with shallow ones. Despite the vast empirical evidence, formal arguments to date are limited and do not capture the kind of networks used in practice. Using tensor factorization, we derive a universal hypothesis space implemented by an arithmetic circuit over functions applied to...
[ Full version ]

Which computer to purchase? Which job to pick? What's the basics of decision making and how technology helps us in hard decision making. More details in the attached poster....
[ Full version ]

...
[ Full version ]

Generalization and simplification are two basic facets of learning. Learning theory gives several mathematical manifestations of these facets. We will present two of these: (i) A generalization model: Probably Approximately Correct learning [Vapnik-Chervonenkis ’71, Valiant ’84], and (ii) A simplification model: Sample compression schemes [Littlestone-Warmuth ’86]. Littlestone and Warmuth have shown that for these mathematical formalizations...
[ Full version ]

Developers and researchers design parallel systems and concurrent algorithms using models---both formal and mental---that abstract away details of the underlying hardware. In practice, however, these hidden hardware/software interactions often have a dominating impact on performance. In this talk, I will show how understanding the interplay between hardware and software can lead to more efficient program execution and better algorithms, which are achieved through more pre...
[ Full version ]

...
[ Full version ]

A method for extended depth of field imaging based on image acquisition through a thin binary phase plate followed by fast automatic computational post-processing is presented. By placing a wavelength dependent optical mask inside the pupil of a conventional camera lens, one acquires a unique response for each of the three main color channels, which adds valuable information that allows blind reconstruction of blurred images without the need of an iterative search process for esti...
[ Full version ]

Tangent vector field processing on discrete surfaces poses many challenges, one of the first being the choice of representation. The popular choice of piecewise-constant vector fields per face leads to very simple expressions and algorithms, at the price of lower accuracy and difficulties in defining derivatives and smoothness energies. Piecewise-linear vector fields on vertices lead to better results, but also to a much higher complexity due to the use of curved triangles, making...
[ Full version ]

We deal with design of models and algorithms for MapReduce. In MapReduce, an input is replicated to several reducers, and hence, the replication of an input dominates the communication cost, which is a performance measure of a MapReduce algorithm. Our MapReduce model is favorable for various matching problems, where the input is replicated as the output. In particular, we present two classes of matching problems, namely the all-to-all and the X-to-Y matching problems, and show tha...
[ Full version ]

I this talk, we will introduce the concept of constraint system decomposition and show how it is beneficial for solving problems in computational geometry. We will also present the geometric and algebraic approaches to constraint decomposition and discuss their advantages and disadvantages with respect to each other....
[ Full version ]

Proof complexity started its life as an attempt to prove that NP is different from coNP. The main goal of the area is to show that no proof system can prove that all negative instances of 3SAT are unsatisfiable using polynomial size proofs. However, this goal has proved too difficult, and nowadays the area concentrates on proving the same for specific proof systems. The focus of this talk is the proof system *cutting planes*, which is the proo...
[ Full version ]

...
[ Full version ]

In this work we consider non-linear classifiers that positively classify a point when it resides in the intersection of $k$ hyperplanes. We learn these classifiers by minimizing the minimax risk of the negative training examples and the sum of hinge-loss of the positive training examples. These classifiers fit typical real-life datasets that consist of a small number of positive data points and a large number of negative data points. Such approach is computationally appealing sinc...
[ Full version ]

You are invited to attend finals of 3rd Intel coding challenge on Wednesday, November 25, 2015, 13:30-17:00, at the EE Meyer Bulding, in lobby 280 and after the break in room 1003. The Intel Coding challenge is open to undergraduate students of Technion CS and EE faculties. The competition is designed to reward contestants who develop innovative, creative, and original applications for PC and tablets. The jury from Intel and Technion would award prizes to 3 b...
[ Full version ]

We continue the study of welfare maximization in unit-demand (matching) markets, in a distributed information model where agent's valuations are unknown to the central planner, and therefore communication is required to determine an efficient allocation. Dobzinski, Nisan and Oren (STOC'14) showed that if the market size is n, then r rounds of interaction (with logarithmic bandwidth) suffice to obtain an n^{1/(r+1)}-approximation to the optimal social welfare. In particular, this i...
[ Full version ]

The talk summarizes the main challenges and opportunities of infrastructure sharing in present and future cellular networks. We begin by assessing, through real-world demand and deployment traces, how sharing can improve the efficiency of present-day networks, especially in rural areas. We then move to next-generation networks, exploring the relationship (and the potential conflict) between sharing and competition regulations. Finally, we study the relationship between trad...
[ Full version ]

In the generalized projection-based M-estimator (gpbM), published a few years ago, multiple structures with different scales were estimated without the scale parameter specified for each inlier structure. However, the gpbM can be defined from much simpler assumptions. From M trials based on the elemental subsets, we find a small group of the data points which produces the minimum sum of the Mahalanobis distances, and expand it to obtain the scale of a structure. This value is used...
[ Full version ]

...
[ Full version ]

We study the relationship between communication and information in 2-party communication protocols when the information is asymmetric. If I_A denotes the number of bits of information revealed by the first party, I_B denotes the information revealed by the second party, and C is the number of bits of communication in the protocol, we show that -- one can simulate the protocol using order I_A + C^{3/4} I_B^{1/4} log C bits of communication, -- one can simulate...
[ Full version ]

Large Computer Clusters are the infrastructure behind The Cloud and Web2 services. These large clusters host highly parallel programs like distributed databases and MapReduce. The parallelism is required for handling the Big Data stored on these clusters. But applications with high degree of parallelism stress the cluster network as they tend to generate correlated traffic bursts of high bandwidth. This stress causes long network latency tail, as well as incast and throughput coll...
[ Full version ]

In these three lectures a (personal) view is given on how from linear/nonlinear objective functions at the input, one can built non robust or robust estimation. If a higher dimensional linear space is considered for nonlinear inputs, one could, directly in total least squares; or iteratively in Levenberg-Marquardt algorithm, non-robustly estimate the features. A robust method of mean shift is presented both for segmentation and tracking. Generalizations, to nonlinear mean shift in...
[ Full version ]

Defining and calculating similarity between objects or images is one of the key topics in computer vision and computer graphics. After reviewing my early work on similarity for shape analysis, I will present my recent work, which focuses on the human EYE. The eye is the only organ in the body where one can non-invasively image blood vessels and nerve tissue. Thus, in addition to detecting eye-specific conditions, by looking into the eye we can diagnose systemic diseases, such as h...
[ Full version ]

The task of parametrizing meshes and, as one important application, quad re-meshing, has been studied extensively in recent years. In this talk I will give an overview of a recent implementation called EvoluteTools T.MAP, which provides mesh parametrization to the wide user community of Rhinoceros 3D. The talk will focus on current and future applications in this user community (among them architects, product designers, naval engineers, and many more), the methods which EvoluteToo...
[ Full version ]

Proof-labeling schemes, introduced by Korman, Kutten and Peleg [PODC 2005], are a mechanism to certify that a network configuration satisfies a given boolean predicate. Such mechanisms find applications in many contexts, for example, the design of fault-tolerant distributed algorithms. In a proof-labeling scheme, predicate verification consists of neighbors exchanging labels, whose contents depends on the predicate. In this paper, we introduce the notion of randomized proof-labeli...
[ Full version ]

In these three lectures a (personal) view is given on how from linear/nonlinear objective functions at the input, one can built non robust or robust estimation. If a higher dimensional linear space is considered for nonlinear inputs, one could, directly in total least squares; or iteratively in Levenberg-Marquardt algorithm, non-robustly estimate the features. A robust method of mean shift is presented both for segmentation and tracking. Generalizations, to nonlinear mean shift in...
[ Full version ]

...
[ Full version ]

Over the last few years we have been developing techniques for analyzing small motions in videos. Our techniques are based on an Eulerian approach for motion processing, which does not explicitly compute motion vectors (as traditionally done in computer vision), but rather analyzes intensity changes in image pixels over time using a combination of spatial and temporal filtering. This lets us turn light measurements efficiently into motion measurements and use ordinary cameras for ...
[ Full version ]

WomenTechPRO Community will host an expert on UX user experience: Dr. Talya Lavie , a leading group of UX \ UI in the HP Software and president of UXPA Israel, will talk about UX engineers and on the science behind the user experience and ergonomics, on Application Lifecycle Management and on what lies ahead in the field Your are all invited. ...
[ Full version ]

We show a directed and robust analogue of a Boolean isoperimetric type theorem of Talagrand. As an application, we give a monotonicity testing algorithm that makes O(\sqrt{n}/\eps^2) non-adaptive queries to a function f, always accepts a monotone function and rejects a function that is \eps-far from being monotone with constant probability. ...
[ Full version ]

Archival data once written are rarely accessed by user, and need to be reliably retained for long periods of time. The challenge of using inexpensive NAND flash to archive cold data was posed recently for saving data center costs. Solid state drives are faster, more power-efficient and mechanically reliable than hard drives (HDs). However, flash of high density is vulnerable to charge leakage over time, and can only be cost-competitive to HD in archival systems if longer re...
[ Full version ]

In these three lectures a (personal) view is given on how from linear/nonlinear objective functions at the input, one can built non robust or robust estimation. If a higher dimensional linear space is considered for nonlinear inputs, one could, directly in total least squares; or iteratively in Levenberg-Marquardt algorithm, non-robustly estimate the features. A robust method of mean shift is presented both for segmentation and tracking. Generalizations, to nonlinear mean shift in...
[ Full version ]

...
[ Full version ]

Please see more details in the attached poster....
[ Full version ]

The watershed algorithm, as proposed by Ch. Lantuéjoul in 1978 and as implemented by L. Vincent and P. Soille in 1991 for image segmentation has been in applications an extremely popular tool ever since, but it is nowadays not so popular at the research level in computer vision. Indeed, whereas it is an efficient algorithm that provides good segmentation results, it is difficult to build new work on this heuristic tool. One of the major issues is that there is some discrepancy be...
[ Full version ]

In view of the recent huge interest in image classification and object recognition problems and the spectacular success of deep learning and random forests in solving these tasks, it seems astonishing that much more modest efforts are being invested into related, and often more difficult, problems of image and multimodal content-based retrieval, and, more generally, similarity assessment in large-scale databases. These problems, arising as primitives in many computer vision tasks,...
[ Full version ]

A geometer, an economist and a Ph.D. student are asked to divide a land-estate fairly among several people. The geometer partitions the land to nicely-shaped parcels - equal-size squares. But the people are unhappy, because squares in different locations have different values (a square near the sea is not like a square in the desert!). The economist partitions the land to parcels with the same economic value, but the people are again unhappy because the plots are all long and narr...
[ Full version ]

In this talk I'll present work on data compression that touches both on the theory of fast compression design and practical aspects as dictated by market and product requirements. I will describe the background, the challenges and our solutions for pushing the time limits of standard compression. No prior knowledge on compression will be assumed.. Bio: Danny Harnik is a researcher in the cloud storage group at IBM Reseach - Haifa. He holds a PhD from The Weizma...
[ Full version ]

We present a method for supervised learning of shape descriptors for shape retrieval applications. Many content- based shape retrieval approaches follow the bag-of-features (BoF) paradigm commonly used in text and image retrieval by first computing local shape descriptors, and then representing them in a `geometric dictionary' using vector quantization. A major drawback of such approaches is that the dictionary is constructed in an unsupervised manner using clustering, unaware of ...
[ Full version ]

Although microscopic processes play a vital role in marine ecosystems and hence, global ecology, the in situ tools to observe them have not existed. Underwater microscopes can play an important role in observing such processes if the difficulties of providing undisturbed measurements can be overcome. In this talk Dr. Jaffe will describe recent work in his lab to improve the available technology to optically image some of the smallest biota in the ocean. This includes a recently de...
[ Full version ]

Internet video streaming already accounts for over 64% all global IP traffic and is projected to increase in the foreseeable future. Consequently, improving the delivery mechanisms of video streaming is an important challenge with real-world consequences that has seen a plethora of protocols and systems been published and made available by industry and academia over the years. Of these systems, peer-to-peer (P2P) delivery, either between client machines, set-top boxes, middle-boxe...
[ Full version ]

CS labs invite you to visit the new project fair on Wednesday, October 21st, 2015, 14:30-16:30 at the Taub Building Lobby. The fair will present the new projects offered by the labs for the winter semester as well as other and activities performed at the labs. You are all invited to come check it out and to join the project's teams....
[ Full version ]

Many efficient data structures use randomness, allowing them to improve upon deterministic ones. Usually, their efficiency and/or correctness are analyzed using probabilistic tools under the assumption that the inputs and queries are independent of the internal randomness of the data structure. In this talk, I'll consider data structures in a more robust model, which we call the adversarial model. Roughly speaking, this model allows an adversary to choose inputs and queries adapti...
[ Full version ]

Modern mobile platforms like Android enable applications to read aggregate power usage on the phone. This information is considered harmless and reading it requires no user permission or notification. We show that by simply reading the phone's aggregate power consumption over a period of a few minutes an application can learn information about the user's location. Aggregate phone power consumption data is extremely noisy due to the multitude of components and applications ...
[ Full version ]

...
[ Full version ]

Facebook is the most popular online social network, with more than a billion active users. As such, Facebook has become a significant source of information and searching within its mass content has become a necessity. Nevertheless, very little research has been done on Facebook Search, due to the difficulty of collecting Facebook data by researchers who are not Facebook employees. In this research we study entity search in Facebook. Queries are public entities, like celebrities,...
[ Full version ]

Solid modelling is a core technology in VLSI process and device modelling. Intel has found that commercial software can be unreliable and slow. An important example is computing an offset of a polyhedron. Prior work on robust computational geometry makes assumptions that conflict with the application requirements, notably the need to perform finite-element analysis on solid models. I will describe preliminary work on a robustness technique that meets these requirements. ...
[ Full version ]

The real time detection of flow anomalies is a critical part of wide range of management and security applications in many Cloud and NFV systems. Solutions that are based on per-flow records become impossible due to the increasing traffic volumes and the limited available resources such as TCAM entries and fast counters. In this paper we study a novel dynamic control mechanism that allows detecting flow anomalies using only a limited number of counters. Starting from the simple...
[ Full version ]

Solid modelling is a core technology in VLSI process and device modelling. Intel has found that commercial software can be unreliable and slow. An important example is computing an offset of a polyhedron. Prior work on robust computational geometry makes assumptions that conflict with the application requirements, notably the need to perform finite-element analysis on solid models. I will describe preliminary work on a robustness technique that meets these requireme...
[ Full version ]

The 4th TCE Summer School on Cyber and Computer Security will be held on Sunday-Wednesday, September 6th-9th, 2015. Location: Room 1003, Meyer Building (EE), Technion, Haifa. Attendance is free, but registration is required Registration and preliminary program available ...
[ Full version ]

Structural biology seeks to determine the structure of proteins and other macromolecular assemblies in order to understand of their function. Ground-breaking advances in detectors’ technology and image processing is revolutionizing this field by enabling the application of electron microscopy for 3D structure determination. In this talk I will present the principles of data acquisition and image processing for cryo-electron microscopy. The talk will focus on the unique propertie...
[ Full version ]

Consider a random constraint satisfaction problem (CSP) over $n$ variables with $m$ constraints, each being a predicate $P$ applied to $k$ random literals. When $m \gg n$ the instance will be unsatisfiable with high probability, and the natural associated algorithmic task is to find a refutation --- i.e., a certificate of unsatisfiability. Understanding the density required for average-case refutation is important in various areas of complexity including cryptography, proof compl...
[ Full version ]

NetFPGA Summer Course 2015 Speakers: Noa Zilberman & Yury Audzevich Host: Mark Silberstein NetFPGA is an open source platform enabling researchers and instructors to build high-speed hardware-accelerated networking systems. The NetFPGA is the de-factor experimental platform for line-rate implementations of network research and it has a family of boards, supporting from 1GE to 100GE. This course will provide an introduction...
[ Full version ]

Transactional memory (TM) eases the task of writing concurrent applications by letting the programmer designate certain code blocks as atomic. One of the main challenges in stating the correctness of TM systems is the need to provide guarantees on the system state observed by live transactions. While several consistency conditions have been suggested for TM, they fall short of formalizing the intuitive semantics of atomic blocks. A TM consistency condition should be weak enough t...
[ Full version ]

On Wednesday. July 29, 2015 we will hold the CYBERDAY workshop on Cyber and Computer Security in the Computer Science department at the Technion together with the Technion computer engineering center. Information, program and registration are available in http://cyberday.org/ The workshop will be held in Hebrew. We will have a POSTER session at the break. Graduate students (and anybody else) wh...
[ Full version ]

...
[ Full version ]

Polynomials over finite fields, and specifically the error correcting codes they induce, have been used extensively in all areas of theoretical computer science. These consist of the evaluation of univariate or multivariate (known as Reed-Solomon(RS) or Reed-Muller(RM) codes, respectively) polynomials on the points of some finite field. One of the limits of using such codes has been that the size of the field limits the degree of polynomials that we can use, and therefore the rate...
[ Full version ]

We introduce a novel method for creating mobile applications, integrating the Android SDK into PlayGo, a scenario-based behavioral programming framework. The framework we implemented allows creating mobile applications simply by using a visual GUI editor, and then incrementally “playing in” scenarios that construct the application behavior. This allows the developer to focus on the application behavior and interface rather than the syntax and code....
[ Full version ]

The total variation (TV) functional is today a fundamental regularizing tool in image processing. It is employed for denoising, deconvolution, optical-flow, tomographic reconstruction, texture and image analysis and more. Recently a spectral TV framework was introduced which extends linear filtering techniques and eigenvalue analysis to a convex nonlinear setting, and specifically to the TV functional. In this work we use the Spectral TV Domain to design advanced texture decompo...
[ Full version ]

Packet classification is a building block in many network services such as routing, filtering, intrusion detection, accounting, monitoring, load-balancing and policy enforcement. Compression has gained attention recently as a way to deal with the expected increase of classifiers size. Typically, compression schemes try to reduce a classifier size while keeping it semantically-equivalent to its original form. Inspired by the advantages of popular compression schemes (e.g. JPEG and ...
[ Full version ]

We consider the random sampling problem in the streaming setting. The algorithm returns, at any time t, an item chosen uniformly at random from all items seen till time t. Our model allows storage for only one item at a time in the stream. We design an algorithm that takes O(log n) random bits and O(log n) space to maintain a random sample at all times. Our algorithm in the streaming setting matches the lower bound on the amount of required randomness for uniform sampling in the o...
[ Full version ]

Initially, SDN has been promoted as a simpler and more flexible way of managing networks than traditional approaches. However, it is unlikely that a single SDN controller vendor will have best-in-class implementation of all network services, so the coexistence of several managing controllers will be a mandatory feature of SDN network management in the near future. Similarly, there is a price for generality and expressiveness to be paid with efficiency of the data plane rep...
[ Full version ]

We present an algorithm for morphing between planar curves (with identical turning number), such that the morph is guaranteed to be a regular homotopy. This means that pinching will not occur in the intermediate curves, or in other words, global intersections are allowed but not local intersections. The algorithm is basically a modification of Sederberg's classical angle-length method necessary for providing the guaranteed regular homotopy. After presenting our results, we will...
[ Full version ]

Conventional cameras record all light falling onto their sensor regardless of the path that light followed to get there. In this talk I will present an emerging family of video cameras that can be programmed to record just a fraction of the light coming from a controllable source, based on the actual 3D path followed. Live video from these cameras offers a very unconventional view of our everyday world in which refraction and scattering can be selectively blocked or enhanced, visu...
[ Full version ]

CS students in the course "Project in Advanced Programming" invite you to an Arduino project fair on Sunday, June 28, 2015, between 13 00-16.30, in room 337 (3rd floor) of CS Taub Building. The fair will exhibit new Arduino pojects designed to facilitate life in all aspects as follows:Please see Hebrew Page...
[ Full version ]

Shape matching is the problem of finding a meaningful transformation between two (or more) shapes that have a priori unknown correspondences. It is the key ingredient for many applications in computer vision, medical imaging and graphics. These applications comprise, for example, shape interpolation, shape retrieval, information transfer and the alignment of scanned data. We propose a method that is based on the physical theory of nonlinear elasticity and is hence a suitable frame...
[ Full version ]

The Technion Computer Engineering Center proudly presents ARM University Day to be held at the Technion on Sunday, June 28th, 2015. This Professor Workshop is held by the ARM University Programme, in partnership with the Technion – Computer Engineering Center, and is aimed at providing Faculty, Post-Graduate members of the Academia, Lab Engineers and Lab Project Supervisors with the knowledge and tools needed to create and run Labs and Project classes based ...
[ Full version ]

CS Software Engineering students invite you to attend the CS Annual Fair in Software Engineering which will be held by the project teams on Thursday-Noon, June 25 2015, between 11:30-13:30, in room 337 (3rd Floor), CS Taub Building as follows: Robotwitter - Respond automatically to hostile campaigns on Twitter ...
[ Full version ]

Google talk on Virtual Reality and Project Tango, and experience the technology first hand. Program: 17:00 - Stepping stones to a software engineering career (for 1st and 2nd year) 18:15 - Virtual reality tech talk and demo...
[ Full version ]

CS students in the course "Project in Advanced Programming" and the course "Project in Computer Communication" invite you to an android applications fair on Wednesday, June 24, 2015, between 14 00-16.00, in room 337 (3rd floor) of CS Taub Building. The fair will exhibit new android applications designed to facilitate life in all aspects and Qualcomm will hold a project competition between the presenting projects: ...
[ Full version ]

Reed-Muller codes encode an m-variate polynomial of degree r by its evaluations over all points in {0,1}^m. The distance of the code is 2^{m-r}, so in the worst case it is impossible to correct more that half that number of errors. For the model of random errors, however, one may hope to achieve better parameters. In this talk we will show an efficient algorithm to correct (roughly) m^{(m-r)/2} random errors for certain ranges of r (coding-theoretic conjectures suggest ...
[ Full version ]

The 2015 Segel-Sudent End of Year Party will be held on Thursday, June 24, 2015, between 12:30-14:30, at the CS Taub Lobby. Free beer, volleyball game with CS Team (on the lawn) plenty of fun are guaranteed. You are all invited!...
[ Full version ]

Biometric authentication is more secure than using regular passwords, as biometrics cannot be "forgotten" and allegedly contain high entropy. Thus, many constructions rely on biometric features for authentication, and use them as a source for "good" cryptographic keys. At the same time, biometric systems carry with them many privacy concerns. Unlike regular passwords, which can be easily changed if compromised, changing biometric traits is far from being easy. Hence, we need to pr...
[ Full version ]

We present a system for solving the holy grail of computer vision -- matching images and text and describing an image by an automatically generated text. Our system is based on combining deep learning tools for images and text, namely Convolutional Neural Networks, word2vec, and Recurrent Neural Networks, with a classical computer vision tool, the Fisher Vector. The Fisher Vector is modified to support hybrid distributions that are a much better fit for the text data. Our method p...
[ Full version ]

In this work, we examine how the support structure in 3D-printing can be optimized, by changing a model's orientation. Specifically, we explore the effect that the orientation of a printed object has on the volume of the needed support structure, directly below the model. We show that the volume of the support is a continuous but non-smooth function, with respect to the orientation angles. We continue by presenting an efficient algorithm, capable of running on a GPU, th...
[ Full version ]

In the setting of secure multiparty computation, a set of parties wish to compute a joint function of their private inputs. The computation should preserve security properties such as privacy, correctness, independence of inputs, fairness and guaranteed output delivery. In the case of no honest majority, fairness and guaranteed output delivery cannot always be obtained. Thus, protocols for secure multiparty computation are typically of two disparate types: protocols that assume an...
[ Full version ]

The Unified Modeling Language (UML) is a widely accepted modeling language for embedded and safety critical systems. As such ,the correct behavior of systems represented as UML models is crucial. Model checking is a successful automated verification technique for checking whether a given system satisfies a desired property. It is widely recognized as an important approach to increase the reliability of hardware and software systems and is vastly used in industry. Unfortunatel...
[ Full version ]

Many architectures for high-performance datacenters have been proposed to meet the ever-growing traffic demands. Surprisingly, recent studies show that even random datacenter topologies outperform much more sophisticated designs, achieving near-optimal throughput and bisection bandwidth, high resiliency to failures, incremental expandability, high cost efficiency, and more. While this highlights the suboptimality of existing datacenters, the inherent unstructuredness and unpredict...
[ Full version ]

...
[ Full version ]

In order to see single photoreceptor cells in the retina, we used hardware and software tools, getting close to the resolution limit, a tenfold improvement. Transparent layers which come in front of the photoreceptors are barely seen by direct imaging, but they play a major role in improving acuity. Radial cells of higher refractive index channel some colours directly into the corresponding photoreceptors to improve their responsivity. Other colours are scattered around, and serve...
[ Full version ]

A Laplacian eigenfunction on a two-dimensional manifold dictates some natural partitions of the manifold; the most apparent one being the well studied nodal domain partition. An alternative partition is revealed by considering a set of distinguished gradient flow lines of the eigenfunction - those which are connected to saddle points. These give rise to Neumann domains (a.k.a Morse-Smale complexes). We define Neumann lines and Neumann domains and present their fundamenta...
[ Full version ]

...
[ Full version ]

Given a human shape, we would like to detect and separate the arms, the head, the legs, but also the fingers and the toes. However, one has to bear in mind that fingers are still part of the arms. This hierarchy can be very simply expressed with the notions of Persistence and Morse-Smale complexes. This talk is in between the previous seminar on Persistence and the next (more advanced) talk of Prof. Rami Band (the same day). To motivate our study, we will first focus o...
[ Full version ]

We are happy to invite you to all times MSc & PhD CS alumni event. The event will be held on Thursday, June 11th at the CS Taub Building, starting at 15:30 and will include a fascinating workshop with famous speakers: Prof. Amnon Shashua, Prof. Hossam Haic, Prof. Michael Aviram and Dr. Haim Shapira and from 19:00, a social and networking happening with guest artist Liior Schuard. program and pre-reg...
[ Full version ]

Vapnik’s fundamental theorem of statistical learning establishes a computational equivalence between optimization (Empirical Risk Minimization) and statistical learning. Is the same true for learning in games? We give a precise answer to this question and describe consequences to the computational complexity of solving zero-sum games with best-response oracles. Joint work with Elad Hazan....
[ Full version ]

We present TeamBoost, a collaborative Multi-Path TCP (MPTCP) solution enabling a client device to leverage its nearby devices over a local network, i.e., WiFi. The client harnesses its nearby devices autonomous connectivity to deliver and relay TCP sub streams of traffic to and from the cloud. The motivation for TeamBoost comes from the need to supply or enhance cloud connectivity to a vehicle infotainment system (the client device). This solution boosts end-to-end performan...
[ Full version ]

Light-field imaging can be scaled up to a very large area, to map the Earth's atmosphere in 3D. Multiview spaceborne instruments suffer low spatio-temporal-angular resolution, and are very expensive and unscalable. We develop sky light-field imaging, by a wide, scalable network of wide-angle cameras looking upwards, which upload their data to the cloud. This new type of imaging-system poses new computational vision and photography problems, some of which generalize prior monocula...
[ Full version ]

Maps of the world around us are used daily for many purposes such as navigation or location-based services. However, our experience with maps is usually limited to a 2D top-down projection. Using a 3D map instead have some advantages, such as easier navigation and orientation, more realistic modeling and simulation of natural phenomena such as water flow or line-of-sight analysis, and better understanding of the elevation differences. In this talk, we will present a few methods t...
[ Full version ]

Given a class of students, and a pool of questions in the domain of study, what subset will constitute a ``good'' exam? Millions of educators are dealing with this difficult problem worldwide, yet the task of composing exams is still performed manually. In this work we present a novel algorithmic framework for exam composition. Our framework requires two input components: a student population represented by a distribution over a set of overlay models, each consisting of a set ...
[ Full version ]

We consider a monopolist seller with n heterogeneous items, facing a single buyer. The buyer has a value for each item drawn independently according to (non-identical) distributions, and his value for a set of items is additive. The seller aims to maximize his revenue. It is known that an optimal mechanism in this setting may be quite complex, requiring randomization and menus of infinite size. Hart and Nisan have initiated a study of two very simple pricing schemes for this setti...
[ Full version ]

Real-time stream processing of Big Data has an increasing demand in modern data centers. There, a continuous torrent of data, created from different streaming data-sources like social networks, video streams, and financial markets, is being processed and analyzed to produce valuable insights about its content, and, in some cases, their value has an expiration date. For example, in a silicon-wafer production inspection system, dozens of high-resolution cameras scan the wafer and ge...
[ Full version ]

Scala is a language that fuses functional and object-oriented programming in a practical package. It interoperates seamlessly with Java and its tools. Scala is now used in a rapidly increasing number of open source projects and companies. It provides the core infrastructure for sites such as Twitter, LinkedIn, Foursquare, Tumblr, and Klout. Send your friends, your students, your colleagues, open to all! No prior Scala knowledge is required. ...
[ Full version ]

In the talk I will present a novel framework for high-throughput cell lineage analysis in time-lapse microscopy images. The proposed algorithm ties together two fundamental aspects of cell lineage construction, namely cell segmentation and tracking, via a Bayesian inference of dynamic models. The proposed contribution exploits the Kalman inference problem by estimating the time-wise cell shape uncertainty in addition to cell trajectory. These inferred cell properties are combined ...
[ Full version ]

We take a new approach to motion planning specifically designed for handling mobile obstacles. Previous methods are modified versions of the algorithms for static obstacles, but with speed-ups designed to ensure real-time operation. Because of this logic, each iteration requires a complete reconstruction of the path, based on the new locations of the robot and the obstacle. Ideally, the algorithm would consist of faster iterations that create a complete approximate path, but can r...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

We construct a general-purpose multi-input functional encryption scheme in the private-key setting. Namely, we construct a scheme where a functional key corresponding to a function f enables a user holding encryptions of x_1, ..., x_t to compute f(x_1, ..., x_t) but nothing else. Our construction assumes any general-purpose private-key single-input scheme (without any additional assumptions), and is proven to be adaptively-secure for any constant number of inputs t. Moreover, it c...
[ Full version ]

We consider a transmission scheme with a single transmitter and multiple receivers over a faulty broadcast channel. For each receiver, the transmitter has a unique infinite stream of packets, and its goal is to deliver them at the highest throughput possible. While such multiple-unicast models are unsolved in general, several network coding based schemes were suggested. In such schemes, the transmitter can either send an uncoded packet, or a coded packet which is a function of a f...
[ Full version ]

We are happy to invite you to the first project fair at CS on Tuesday, May 26 2015, between 15:30-17:30 at the CS Taub lobby, The event will host top executives from the hi-tech industry. The participating students will get a great opportunity to present their projects accomplished throughout their studies. Furthermore, the firms representatives will be exposed to the applicable aspects of the degree. and to the students' development and implementing abilities...
[ Full version ]

Our industry has reaped the benefits of Moore's Law for 50 years now, making integrated circuits that have grown from tens to billions of transistors and performing an increased range of functions from memory to logic to signal processing. Scaled transistors have provided significant improvements in performance and low power, but the main benefit of scaling has been lower cost per transistor. As we scale to 10 nm and below it is becoming increasingly difficult to achieve t...
[ Full version ]

We will give a new derivation of the Bloch-Torrey equation that governs the behaviour of water molecules in porous media and biological tissues. We will show that this new derivation reveals new term that was ignored in the past. We further show how one can solve this equation analytically for few specific geometries. These solutions enables us to solve the inverse problem of finding the geometric parameters at a micron level from NMR measurements. These predictions ar...
[ Full version ]

Isogeometric analysis (IGA) has brought new problems to the field of Computer Aided Geometry for Design (CAGD); in return the latter brings to the computation of solution of PDEs new methods and algorithms. The present work develops both aspects. In the first one, IGA has renewed the interest for trivariate spline objects and their manipulation: construction of multi patch primitives that are C0 or G1, construction of trivariate splines resulting from classic CSG operations such a...
[ Full version ]

In this presentation we demonstrate the installation of persistent firmware modifications into the EFI boot ROM of Apple's popular MacBooks. The bootkit can be easily installed by an evil-maid via the externally accessible Thunderbolt ports and can survive reinstallation of OSX as well as hard drive replacements. Once installed, it can prevent software attempts to remove it and could spread virally across air-gaps by infecting additional Thunderbolt devices. Bio: Tr...
[ Full version ]

The gold standard for a classifier is the condition of optimality attained by the Bayesian classifier. Within a Bayesian paradigm, if we are allowed to compare the testing sample with only a single point in the feature space from each class, the optimal Bayesian strategy would be to achieve this based on the (Mahalanobis) distance from the corresponding means. The reader should observe that, in this context, the mean, in one sense, is the most central point in the respective distr...
[ Full version ]

I will survey developments in the application of invariants of various types, including invariant histograms and differential invariant signatures, for object recognition and symmetry detection in images. I will present some recent progress, including applications to automated jigsaw puzzle assembly, in both 2D and 3D, as well as detection of melanomas and cancerous tumors...
[ Full version ]

Proofs of proximity are probabilistic proof systems in which the verifier only queries a sub-linear number of input bits, and soundness only means that, with high probability, the input is close to an accepting input. In their minimal form, called Merlin-Arthur proofs of proximity (MAP), the verifier receives, in addition to query access to the input, also free access to an explicitly given short (sub-linear) proof. A more general notion is that of an interactive proof of proximit...
[ Full version ]

Future cellular networks will be owned by multiple parties, e.g., two mobile operators, each of which controls some elements of the access and backhaul infrastructure. In this context, it is important that as much traffic as possible is processed by the same party that generates it, i.e., that the coupling between traffic and network ownership is maximized. Software-defined backhaul networks can attain this goal; however, existing management schemes ignore ownership alt...
[ Full version ]

Parallel machine scheduling has been extensively studied in the past decades, with applications ranging from production planning to job processing in large computing clusters. We study some of these fundamental optimization problems, as well as their reoptimization variants. In this talk, we present improved bounds for job scheduling on unrelated parallel machines, with the objective of minimizing the latest completion time (or, makespan) of the schedule. We consider the subclas...
[ Full version ]

The classical method of moving frames was developed by Elie Cartan into a powerful tool for studying the geometry of submanifolds under certain geometrical transformation groups. In this talk, I will present a new foundation for moving frame theory based on equivariant maps. The method is completely algorithmic, and applies to very general Lie group actions and even infinite-dimensional pseudo-groups. It has led to a wide variety of new applications, ranging over classical differe...
[ Full version ]

We propose a general scheme for detecting critical locations (of dimension zero) of piecewise polynomial multivariate equation systems. Our approach generalizes previously known methods for locating tangency events or self-intersections, in contexts such as surface-surface intersection (SSI) problems and the problem of tracing implicit plane curves. Given the algebraic constraints of the original problem, we formulate additional constraints, seeking locations where the ...
[ Full version ]

We define a new class of semi-definite programs in which the input arrives online. We use the primal dual method to develop an algorithm for this setting of semi-definite programs while maintaining the monotonicity requirements of an online algorithm. Our algorithm achieves logarithmic competitive ratios for both the primal and dual problems. We will discuss applications from quantum information theory, such as the Quantum Hypergraph Covering problem, and also observe the behavior...
[ Full version ]

A tech lecture by Aharon Aharon from Apple will be held on Wednesday, May 13 2015, at 12:30, in Auditorium 1003 (10th floor), EE Meyer Building. You are all welcome....
[ Full version ]

An error-correcting code $C \subseteq \F^n$ is called $(q,\epsilon)$-strong locally testable code (LTC) if there exists a tester that makes at most $q$ queries to the input word. This tester accepts all codewords with probability 1 and rejects all non-codewords $x\notin C$ with probability at least $\epsilon \cdot \delta(x,C)$, where $\delta(x,C)$ denotes the relative Hamming distance between the word $x$ and the code $C$. The parameter $q$ is called the query complexity and the p...
[ Full version ]

Online infrastructure often depends on security-critical authorities such as logging, time, and certificate services. Authorities, however, are vulnerable to the compromise of one or a few centralized hosts yielding "weakest-link" security. We propose collective authorities or cothorities, an architecture enabling thousands of participants to witness, validate, and co-sign an authority's public actions, with moderate delays and costs. Hosts comprising a cothority form an efficient...
[ Full version ]

The massive parallelism and resource sharing embodying today’s cloud business model not only exacerbate the security challenge of timing channels, but also undermine the viability of defenses based on resource partitioning. This paper proposes hypervisor-enforced timing mitigation to control timing channels in cloud environments. This approach closes “reference clocks” internal to the cloud by imposing a deterministic view of time on guest code, and uses timing mitigators to...
[ Full version ]

Many people have legitimate needs to avoid their online activities being tracked and linked to their real-world identities - from citizens of authoritarian regimes, to everyday victims of domestic abuse or law enforcement officers investigating organized crime. Current state-of-the-art anonymous communication systems are based on onion routing, an approach effective against localized adversaries with a limited ability to monitor or tamper with network traffic. In an environment ...
[ Full version ]

Recent advances in realtime performance capture and virtual reality have brought within reach a new form of human communication. Capturing dynamic facial expressions of a user and retargeting these expressions to a digital character in realtime allows enacting arbitrary virtual avatars with live feedback. Compared to communication via recorded video streams that only offer limited ability to alter one’s appearance, such technology opens the door to fascinating new applications i...
[ Full version ]

Memory is a large component of computer system cost and current trends indicate this cost is increasing as a fraction of the total. Emerging applications such as in-memory databases, virtual machines, and big-data key-value stores demand more memory relative to compute. Some of these high-memory applications incidentally store many duplicate values in memory: in some cases, duplicates account for over 75% of the total. This talk describes a practical deduplicated memory that is co...
[ Full version ]

Abstract: Software Watermarking is the process of transforming a program into a functionally equivalent “marked” program in such a way that it is computationally hard to remove the mark without destroying functionality. Barak, Goldreich, Impagliazzo, Rudich, Sahai, Vadhan and Yang (CRYPTO 2001) defined software watermarking and showed that the existence of indistinguishability obfuscation implies that software watermarking is impossible. Given the recent candidate construction...
[ Full version ]

Weakly secure codes aim to hide information about individual packets as well as small groups of packets from an eavesdropper that can intercept wireless transmissions or has access to a small number of storage nodes. Such codes are a practical alternative to traditional information-theoretic schemes that hide information about the entire set of files from the eavesdropper. The weakly secure codes do not use random keys, and as a result have better performance and lower overhead th...
[ Full version ]

...
[ Full version ]

Solid sweep is the envelope of the swept volume of a given solid moving along a one parameter family of rigid motions in 3 dimensional space. Solid sweep is a powerful and versatile surface design tool which has found uses in many areas like CNC-machining verification, collision detection and packaging. This work aims to develop a robust computational framework for sweeps in the realm of boundary representation (brep) format, wherein, the input solid as well as the outputted swe...
[ Full version ]

Conventional magnetic recording media are composed of basic magnetizable two-dimensional units called grains, which might be random in size and shape. Recently, a new technological enhancement was proposed, which enables magnetizing areas as small as the size of grains. This novelty effectively created a different type of medium, in which one observes a new type of errors. Handling such errors is the main subject of this work. We first consider a combinatorial model of this mediu...
[ Full version ]

You are invited to the final stage of the 2014 Amdocs Best Project Contest. The competing teams will present and talk about their projects. The event will take place on Wednesday, April 29, 2015, 16:30-18:30, in Auditorium 2, CS Taub Building. You are all invited to cheer, support and be exposed to outstanding projects....
[ Full version ]

CS invites you to a STARTUP FAIR and recruitment by the presenting firms: Click-tale, Forter, Kamimnario, Ravello, Outbrain, Taxi, Wix.com and more. In addition, lectures will be given by the firms representatives. The event will take place on Wednesday, April 29, 2016, between 12:30-14:30 at the CS Taub Lobby. The lectures will be held in Auditorium 2. You are all welcome! ...
[ Full version ]

We prove that any graph excluding $K_r$ as a minor can be partitioned into clusters of diameter at most $\Delta$ while removing at most $O(r/\Delta)$ fraction of the edges. This improves over the results of Fakcharoenphol and Talwar, who building on the work of Klein, Plotkin and Rao gave a partitioning that required to remove $O(r^2/\Delta)$ fraction of the edges. Our result is obtained by a new approach to relate the topological properties (excluding a minor) of a gr...
[ Full version ]

Most modern computer systems use demand paging based virtual memory. Address translation for virtual memory support in I/O devices is known to be useful. It allows SR-IOV (single root I/O virtualization) direct device assignment in virtualization scenarios and kernel bypassing access to I/O devices by user space applications. However, these mechanisms require memory pinning and strict, static memory allocation as they do not support demand paging. This deficiency prevents memory ...
[ Full version ]

Log-structured data stores (LSM-DSs) are widely accepted as the state-of-the-art implementation of key-value stores. They replace random disk writes with sequential I/O, by accumulating large batches of updates in an in-memory data structure and merging it with the on-disk store in the background. While LSM-DS implementations proved to be highly successful at masking the I/O bottleneck, scaling them up on multicore CPUs remains a challenge. This is nontrivial due to their often ri...
[ Full version ]

Predictions in modern inference problems can be increasingly understood in terms of high-dimensional structures such as arrangements of objects in computer vision, parses in natural language processing or molecular structures in computational biology. To interactively annotate an image, one needs to efficiently sample possible interactions to suggest the user to annotate areas with high uncertainty. Unfortunately, traditional approaches to handle high-dimensional structures involv...
[ Full version ]

In this talk, I will describe a joint Microsoft Research and Bing project to study the possibility of using field programmable gate arrays to accelerate cloud applications. We developed an FPGA card that plugs into a Microsoft designed cloud server and used it to accelerate a significant portion of Bing’s search engine. Because the application does not fit into a single FPGA, we spread the application across multiple FPGAs spread across multiple servers and connected together wi...
[ Full version ]

Several interesting observations about the Human Visual System have been investigated in recent years: 1. Sensory data is encoded in the visual cortex using color opponent space (luminance / chrominance space). 2. The distribution of RGB cones over the retina is non-uniform and highly biased towards the red/green cones. 3. The Human Visual System has high spatial sensitivity in the luminance band and low spatial sensitivity in the chrominance bands. These observations have been vi...
[ Full version ]

In this talk we show how logic can be eliminated from several meta-theorems relating the definability of graph parameters in sublogics of Second Order Logic to their computational complexity over certain graph classes. A famous example of such a meta-theorem is Courcelle's theorem, stating that graph parameters definable in Monadic Second Order Logic are computable in polynomial time over graph classes of bounded tree-width. A logic-free version of this theorem was...
[ Full version ]

It is naturally easier to comprehend simple code relative to complicated code. We suggest that code regularity — where the same structures are repeated time after time — may significantly reduce complexity, because once one figures out the basic repeated element it is easier to understand additional instances. We demonstrate this by controlled experiments where subjects perform cognitive tasks on different versions of the same basic function. The results also suggest a cont...
[ Full version ]

The 2015 open day invite outstanding undergraduates from all universities to come to the Technion and learn about the faculties of Electrical Engineering and Computer Science, meet faculty and graduate students and hear a fascinating talk by Dr. Aya Sofer, Director at IBM Research - Haifa, on graduate degrees from the perspective of the high-tech industry. The event will be held on Wednesday, April 15 2015, between 10:30-16:30, in CS Taub Building, room 337 (3...
[ Full version ]

...
[ Full version ]

This work is a theoretical study of the following Non-Rigid Structure from Motion problem. What can be computed from a monocular view of a parametrically deforming set of points? We treat various variations of this problem for affine and polynomial deformations with calibrated and uncalibrated cameras. We show that in general at least three images with quasi-identical two deformations are needed in order to have a finite set of solutions of the points' structure and calculat...
[ Full version ]

The amount of programming tutorial videos on the web increases on a daily basis. Video hosting sites such as YouTube host millions of video lectures, with many programming tutorials for various languages and platforms. Automatically understanding the content of such videos is desirable for many purposes, including search, targeting of ads, and referrals to semantically related content. We present a novel approach for extracting code from videos. Our technique extracts and recogni...
[ Full version ]

Induction algorithms have steadily improved over the years, resulting in powerful methods for learning. However, these methods are constrained to use knowledge within the supplied feature vectors. Recently, a large collection of common-sense and domain specific relational knowledge bases have become available on the web. The natural question is how these knowledge bases can be exploited by existing induction algorithms. In this work we propose a novel algorithm for using relation...
[ Full version ]

There is a growing, vital need for fast algorithms for problems that are unlikely to admit efficient solutions, based on classical computational complexity theory. Parameterized Complexity is an exciting paradigm for coping with computationally hard problems, which is amazingly doable mathematically on a routine basis. In a nutshell, this paradigm aims to reduce the running times of algorithms for NP-hard problems, by confining the combinatorial explosion to a parameter k. In th...
[ Full version ]

In the near future, cloud providers will sell their users virtual machines with CPU, memory, network, and storage resources whose prices constantly change according to market-driven supply and demand conditions. Running traditional operating systems in these virtual machines is a poor fit: traditional operating systems are not aware of changing resource prices and their sole aim is to maximize performance with no consideration for costs. Consequently, they yield low profits. We ...
[ Full version ]

IOMMUs provided by modern hardware allow the OS to enforce memory protection controls on the DMA operations of its I/O devices. An IOMMU translation management design must scalably handle frequent concurrent updates of IOMMU translations made by multiple cores, as occur in high throughput I/O workloads. Today, however, OSes experience performance meltdowns under these workloads. In this talk we will explore scalable IOMMU management designs, which would address the two main bot...
[ Full version ]

Intel Software Guard Extensions (Intel SGX) is an extension to Intel Architecture designed to increase the security of software. In this approach, rather than attempting to identify and isolate all the malware on the platform, legitimate software can be sealed inside an enclave and protected from attack by malware, irrespective of the its privilege level. In the talk I will touch on the building blocks of security, describe the basics of Intel SGX, and show how the components, com...
[ Full version ]

...
[ Full version ]

The popularity of low-cost RGB-D scanners is increasing on a daily basis. Nevertheless, existing scanners often cannot capture subtle details in the environment. We present a novel method to enhance the depth map by fusing the intensity and depth information to create more detailed range profiles. The lighting model we use can handle natural scene illumination. It is integrated in a shape from shading like technique to improve the visual fidelity of the reconstructed object. Unlik...
[ Full version ]

The problem of planar mapping and deformation is central in computer graphics, where applications range from image warping and character animation, to non-rigid registration and shape analysis. I will present a framework for adapting general, smooth, function bases for constructing provably good planar mappings. The term "good" in this context means the map has no fold-overs (injective), is smooth, and has low isometric or conformal distortion. Our approach bridges the gap between...
[ Full version ]

You are kindly invited to the 2015 Yahoo! Labs Israel Technion Grad Student Event on Sunday March 29, 2015, Taub Building, Technion Campus as follows: 14:30-15:00 at CS Taub, Auditorium 2 Welcome by Prof. Irad Yavneh, CS Dean Welcome by Dr. Yoelle Maarek, VP Research, Yahoo! Labs: “Why do you need a graduate degree in leading Web companies” 15:00-15:30 at CS Taub, Auditorium 2 Research Talk by Prof. Eugene Agichtein: “...
[ Full version ]

NVIDIA GPUs are powering a revolution in machine learning. With the rise of deep learning algorithms, in particular deep convolutional neural networks, computers are learning to see, hear, and understand the world around us in ways never before possible. Image recognition and detection systems are getting close to and in some cases surpassing human-level performance. I will talk about deep learning in the context of several new NVIDIA initiatives ranging from hardware platforms, s...
[ Full version ]

You are invited to HacktonPro2015 from CS and WomenTechPRO. The event will take place on Thursday-Friday, March 26-27 2015, at the CS Taub Building, Technion and will spread on 24 hours of creation and development of the most creative ideas with the help of expert mentors in all aspects and technologies and from all over the country and the leading...
[ Full version ]

We introduce and construct a pseudo-random object which we call a local correlation breaker (LCB). Informally speaking, an LCB uses a weak-source of randomness to "break" local correlations between random variables. As an application of LCBs, we construct a 3-source extractor for entropies delta*n, O(log n) and O(loglog n), for any constant delta. . No prior knowledge is assumed....
[ Full version ]

A short overview of the NSF Variability Expedition will be given, followed by an overview of a particular result: SmartBalance. Due to increased demand for higher performance and better energy efficiency, MPSoCs are deploying heterogeneous architectures with architecturally differentiated core types. However, the traditional Linux-based operating system is unable to exploit this heterogeneity since existing kernel load balancing and scheduling approaches lack support fo...
[ Full version ]

...
[ Full version ]

We propose an approach to reduce the runtime of Exemplar-SVM when multiple detectors are to be applied to the same image. Our underlying assumption is that image windows of similar appearance will yield similar detection responses. Therefore, rather than applying all the detectors to all the candidate image windows, we compute the detector scores only once, for a finite pool of image windows. The detection score for each candidate is constructed from these values, and serves as an...
[ Full version ]

Shapes are usually described by their geometrical characteristics. Obviously, a rabbit doesn't look like a horse, because of its curvature for instance. However, some characteristics of a shape are independent of its geometry. A rugby ball is still a ball, and a doughnut looks like a tire. Those are called topological features of the shape, and homology is one the tool that mathematicians use to separate a sphere from a torus. A rabbit being equivalent to a horse, this often makes...
[ Full version ]

Transfer learning techniques are concerned with the creation of high performance predictive models, challenged with sparsely labeled training examples and using related learning tasks for which sufficient training sets are available. Transfer learning can be motivated by a common scenario in which we obtain a large annotated training set for the problem at hand ("source") and use it to build a classifier, only to learn that the examples came from a related, but different pr...
[ Full version ]

Code similarity is a central challenge in many programming related applications, such as code search, automatic translation, and programming education. We present a novel approach for establishing the similarity of code fragments by computing textual similarity between their corresponding textual descriptions. In order to find textual descriptions for code fragments, we leverage the collective knowledge captured in question answering sites, blog posts and other sources. Because...
[ Full version ]

We study the topological properties of wireless communication maps and their usability in algorithmic design. We consider the SINR model, which compares the received power of a signal at a receiver against the sum of strengths of other interfering signals plus background noise. To model the reception regions, we use the convenient representation of an \emph{SINR diagram}, introduced in \cite{Avin10journal}, which partitions the plane into $n$ reception zones, one per station, and ...
[ Full version ]

I will start with a quick background on wireless sensor networks (WSNs), and some of the security challenges created by their unique features. Random key pre-distribution schemes address some of the difficulties by randomly assigning secure keys to sensor nodes prior to network deployment. The discussion will focus on two very different schemes, namely the scheme of Eschenauer and Gligor, and the random pairwise scheme of Chan et al. Each of these schemes induces a non-...
[ Full version ]

Online learning and competitive analysis are two widely studied frameworks for online decision-making settings. Despite the frequent similarity of the problems they study, there are significant differences in their assumptions, goals and techniques, hindering a unified analysis and richer interplay between the two. In this research we provide several contributions in this direction. First, we provide a single unified algorithm which by parameter tuning, interpolates between optim...
[ Full version ]

A discrete model of the Ricci flow for images is presented, using a purely combinatorial method for calculating the Ricci curvature, based on R. Forman's work on generalized Laplacians for cell complexes. The adaptation of Forman's curvature function for cell complexes to images is natural and straightforward, using geometric properties of the image as a surface embedded in Euclidean space. It is shown that the framework presented is far more applicable to images than other existi...
[ Full version ]

N-PolyVector fields are introduced, which are a generalization of N-RoSy fields for which the vectors are neither necessarily orthogonal nor rotationally symmetric. A novel representation for N-PolyVectors as the root sets of complex polynomials is given, in addition to the analysis of their topological and geometric properties. A smooth N-PolyVector field can be efficiently generated by solving a sparse linear system without integer variables. This flexibility of N-PolyVector fie...
[ Full version ]

Shared-memory multiprocessors execute concurrently multiple threads of computation that communicate and synchronize through shared memory. Typically, this communication and synchronization is done via concurrent data structures. The efficiency of these data structures is crucial to performance. Furthermore, new challenges arise in designing scalable concurrent data structures that can perform well with an increasing number of concurrent threads. Non-blocking data structures are s...
[ Full version ]

This work aims to design and implement a software/hardware bundle that could serve as a technical sailing trainer for the sailors of the Laser dinghy sailboat (or single-handed, main sail boats). This trainer should allow sailors, especially in the youth level, to practice in a genuine environment even when a human trainer is not available. Trying to keep the trainer simple and affordable, the bundle is composed of a smartphone and a small set of Wi-Fi enabled cameras placed in...
[ Full version ]

...
[ Full version ]

This work presents a novel approach for detecting inliers in a given set of correspondences (matches). It does so without explicitly identifying any consensus set, based on a method for inlier rate estimation (IRE). Given such an estimator for the inlier rate, we also present an algorithm that detects a globally optimal transformation. We provide a theoretical analysis of the IRE method using a stochastic generative model on the continuous spaces of matches and transformations. Th...
[ Full version ]

The reconstruction of the 3D world from images is among the central challenges in computer vision. Starting in the 2000s, researchers have pioneered algorithms which can reconstruct camera motion and sparse feature-points in real-time. In my talk, I will show that one can autonomously fly quadrotors and reconstruct their environment using onboard color or RGB-D cameras. In particular, I will introduce spatially dense methods for camera tracking and reconstruction which do not r...
[ Full version ]

Advance reservation (AR) services form a pillar of many branches of the economy, including transportation, lodging, dining, and, more recently, cloud computing. In this work, we use game theory to analyze an AR system in which customers differ in their lead (arrival) times and the number of customers requesting service is random. Based on statistical information, the customers decide whether or not making an advance reservation of server resources for a fee. We prove that only ...
[ Full version ]

We present a static analysis for mining synchronization patterns from a large number of code samples. Our analysis uses automata-based abstraction to capture sequences of API operations and synchronization operations. By combining information about the sequence of API calls with synchronization operations, we uncover the intended synchronization based on the semantics of the underlying component. We address both lock-based synchronization and lock-free patterns. Our approach is i...
[ Full version ]

We consider a distributed system of $n$ nodes communicating in the synchronous Vertex-Congest model. In this model, in each round, each node sends the same message of size $O(\log n)$ to all neighbors. We investigate the problem of information spreading, where each node has an initial token, and the goal is to collect tokens of all other nodes. Focusing on communication graphs that are $k$-vertex-connected, we argue that since the existing near-optimal algorithm that requires $O(...
[ Full version ]

The work addresses the two main challenges that have so far prevented adaption of chirp transmission in medical ultrasound, despite its potential to achieve a significantly enhanced SNR. The first is associated with the high sidelobe level resulting from signal compression using a matched filter. The second refers to the level of the sidelobes that is further increased since the pulse shape is distorted in the biological tissue in an unknown manner. This distortion is a result of ...
[ Full version ]

Various problems involving image and shape representation, analysis and processing share the following common denominator: the complexity and the accuracy of the solution depend on the specific problem formulation and data representation being used. We propose to utilize geometric problem formulations together with novel data representation domains for efficient object segmentation in images and three dimensional shape matching. For automatic image segmentation, we consider the ...
[ Full version ]

This talk surveys several state of the art techniques to approximate sets and multi sets efficiently. While Bloom filters and their variants are well researched, this talk argues that compact hash table encoding is a better approach. To back up this claim, we introduce two novel data structures that are more space efficient than previously suggested alternatives. The talk will also present novel applications to approximate multi set representation in the field of cache management ...
[ Full version ]

We consider flexible resource allocation problems in networks and cloud services, where the goal is to optimally utilize a limited amount of a resource (e.g., cloud servers, bandwidth, or storage capacity) available along a given time interval. Such practical scenarios give rise to variants of classic scheduling or packing problems, whose solutions lead to better performance in such criteria as throughput, revenue and resource utilization. We study the theoretical as well as pract...
[ Full version ]

For decades technology scaling has decreased the cost of computing to the point where it can be included in almost anything. We have become accustomed to computing becoming faster, cheaper, and lower power, so we simply assume it will continue. While scaling has never been easy, a number of factors have made scaling increasingly difficult this past decade, and have caused power to become the principal constraint on performance. To continue scaling, new technologies (e.g., memristo...
[ Full version ]

The local self-similarity (LSS) image prior provides a reliable and efficient source for example patches in natural images. We present a new image upscaling method in which the LSS prior is: (i) fulfilled exactly along image singularities, e.g., edges, and (ii) its use is maximized where it holds and suppressed otherwise. The new approach interleaves the pixels of several pre-scaled images, which differ by sub-pixel offsets, to form the low-frequency layer of the output image. The...
[ Full version ]

The concurrent memory reclamation problem is that of devising techniques to allow a deallocating thread to verify that no other concurrent threads, in particular ones executing read-only traversals, have a reference to the block being deallocated. To date, there is no satisfactory solution to this problem: existing tracking methods like hazard pointers, reference counters, or epoch-based RCU, are either prohibitively expensive or require significant programming expertise, to the e...
[ Full version ]

With the advent of easier to parse languages such as Java, and the availability on the Internet of open-source software repositories, complete with versioning histories, empirical studies at scale of software engineering metrics and measurements have become possible and feasible. We take up the questions of if and how "structured GOTO" statements impact defect proneness, and of which what concept of size yields a superior metric for defect prediction. We view the topic thro...
[ Full version ]

As data becomes dynamic, large, and distributed, there is increasing demand for what have become known as distributed stream algorithms. Since continuously collecting the data to a central server and processing it there incurs very high communication and computation complexities, it is desirable to define local conditions at the nodes, such that - as long as they are maintained - some desirable global condition holds. A generic algorithm which proved very useful for reducing commu...
[ Full version ]

You are invited to visit the renewed SSDL (Systems and Software Development Laboratory) and in the Xbox hall. The opening will be on Thursday, January 22, between 12:30-13:00, at the second floor of Taub Building. The inauguration ceremony, in cooperation with Microsoft, and project presentation by the Arduino course, will be held at 13:30, in room 337 (3rd floor). You are all i...
[ Full version ]

How does Google provide instant answers to user queries such as: "What's the weather tomorrow?", "Who won the Miami Heat game last night?" or "Apple stock price"? Come to our technical talk about Google Live Results on Wednesday, January 21 at 12:30 in classroom 7 Taub Building. Google Live Results use natural language processing and Google's Knowledge Graph to understand user questions in numerous languages and to provide text a...
[ Full version ]

We consider the task of multiparty computation performed over networks in the presence of random noise. Given an $n$-party protocol that takes $R$ rounds assuming noiseless communication, the goal is to find a coding scheme that takes $R'$ rounds and computes the same function with high probability even when the communication is noisy, while maintaining a constant asymptotical rate, i.e. while keeping $R/R'$ positive as $n$ and $R$ increase. Rajagopalan and Schu...
[ Full version ]

Two problems with parallel summation of floating-point numbers on GPUs are loss of precision and non-reproducible results. The precision loss is due to round-off error propagation, and the lack of bit-accurate consistency across platforms and setups can be attributed to the non-associative nature of floating-point addition. To address these problems, we implemented a new method for efficient bit-accurate parallel summation of double-precision numbers on GPUs. This metho...
[ Full version ]

In automata theory, a Hankel matrix $H(f,\circ)$ is an infinite matrix where the rows and columns are labeled with words $w$ over a fixed alphabet $\Sigma$, and the entry $H(f,\circ)_{u,v}$ is given by $f(u \circ v)$, where $f$ is a real-valued word function and $\circ$ denotes concatenation. A classical result of G.W. Carlyle and A. Paz (1971) in automata theory characterizes real-valued word functions $f$ recognizable by weighted automata as the functions for which the Hanke...
[ Full version ]

Blind people face challenges when using mainstream technology that relies heavily on visual feedback. Consider your mobile phone as an example. How would you interact with it without looking at the screen? In my talk, I’ll present the state-of-the-art in access technology and discuss social issues involved in its design. I’ll then discuss several of my research projects, including new methods and studies that explore how blind people can enter text efficiently on mobile device...
[ Full version ]

...
[ Full version ]

The Lucas-Kanade (LK) method is a classic tracking algorithm exploiting target structural constraints thorough template matching. Extended Lucas Kanade or ELK casts the original LK algorithm as a maximum likelihood optimization and then extends it by considering pixel object / background likelihoods in the optimization. Template matching and pixel-based object / background segregation are tied together by a unified Bayesian framework. In this framework two log-likelihood ter...
[ Full version ]

...
[ Full version ]

Voronoi diagrams are one of the most powerful data structures in computational geometry. Barequet, Dickerson, and Goodrich [D&CG 2001] introduced the notion of polygon-offset distance function for points. The polygon-offset distance function is more natural than scaling and is useful to model many problems which arise in manufacturing tolerances. Using this distance function, they studied the nearest- and farthest-site Voronoi diagrams of point sites in the plane. In this talk, I ...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

NAND flash, used in modern SSDs, is a write once medium, where each memory cell must be erased prior to writing. The lifetime of an SSD is limited by the number of erasures allowed on each cell. Thus, minimizing erasures is a key objective in SSD design. A promising approach to eliminate erasures and extend SSD lifetime is to use Write Once Memory (WOM) codes, designed to accommodate additional writes on write once media. However, these codes inflate the physically stor...
[ Full version ]

Metric data structures (distance oracles, distance labeling schemes, routing schemes) and low-distortion embeddings provide a powerful algorithmic methodology, which has been successfully applied for approximation algorithms [LLR95], online algorithms [BBMN11], distributed algorithms [KKMPT12] and for computing sparsifiers [ST04]. However, this methodology appears to have a limitation: the worst-case performance inherently depends on the cardinality of the metric, and one could ...
[ Full version ]

...
[ Full version ]

We present a deep layered architecture that generalizes classical convolutional neural networks (ConvNets). The architecture, called SimNets, is driven by two operators, one being a similarity function whose family contains the convolution operator used in ConvNets, and the other is a new soft max-min-mean operator called MEX that realizes classical operators like ReLU and max pooling, but has additional capabilities that make SimNets a powerful generalization of ConvNets. Three...
[ Full version ]

...
[ Full version ]

...
[ Full version ]

As awareness for privacy of Domain Name System (DNS) is increasing, a number of mechanisms for encryption of DNS packets were proposed. We study the prominent defences, focusing on the privacy guarantees, interoperability with the DNS infrastructure, and the efficiency overhead. In particular: - We explore dependencies in DNS and show techniques that utilise DNS specific side channel leaks allowing to infer information about the target domain in an encrypted DNS packe...
[ Full version ]

...
[ Full version ]