Colloquia and Seminars
To join the email distribution list of the cs colloquia, please visit the list subscription page.
Computer Science events calendar in HTTP ICS format for of Google calendars, and for Outlook.
Academic Calendar at Technion site.
Upcoming Colloquia & Seminars
Roded Zats, M.Sc. Thesis Seminar
Advisor: Prof. Roy Schwartz
We consider the family of Correlation Clustering optimization problems under fairness constraints.
In Correlation Clustering we are given a graph whose every edge is labeled either with a $+$ or a $-$, and the goal is to find a clustering that agrees the most with the labels: $+$ edges within clusters and $-$ edges across clusters.
The notion of fairness implies that there is no over, or under, representation of vertices in the clustering: every vertex has a color and the distribution of colors within each cluster is required to be the same as the distribution of colors in the input graph.
Previously, approximation algorithms were known only for fair disagreement minimization in complete unweighted graphs.
We prove the following:
$(1)$ there is no finite approximation for fair disagreement minimization in general graphs unless $ P=NP$ (this hardness holds also for bicriteria algorithms); and $(2)$ fair agreement maximization in general graphs admits a bicriteria approximation of $\approx 0.591$ (an improved $\approx 0.609$ true approximation is given for the special case of two uniformly distributed colors).
Roy Benjamin, M.Sc. Thesis Seminar
Advisor: Dr. Kira Radinsky
Recent global events have emphasized the importance of accelerating the drug discovery process. This process may take more than a decade and its overall cost might exceed one billion dollars. A way to deal with these issues is to use machine learning to increase the rate at which drugs are made available to the public while reducing the cost of the entire process. However, chemical labeled data for real-world applications is extremely scarce making traditional approaches less effective.
A promising course of action for this challenge is to pretrain a model using related tasks with large enough datasets, with the next step being finetuning it for the desired task. This is challenging as creating these datasets requires labeled data or expert knowledge.
To aid in solving this pressing issue, in this thesis we introduce MISU - Molecular Inherent SUpervision, a unique method for pretraining graph neural networks for molecular property prediction.
Our method leapfrogs past the need for labeled data or any expert knowledge by introducing three innovative components that utilize inherent properties of molecular graphs to induce information extraction at different scales, from the local neighborhood of an atom to substructures in the entire molecule.
We evaluate our framework on six chemical property prediction tasks. We show that our method reaches state-of-the-art results compared to numerous baselines. We conduct a thorough ablation experiment and emphasize the contribution of each component in the method. In addition, we explore the effect of MISU on various GNN architectures and find our method is consistent with work done on supervised pretraining.