Colloquia and Seminars

Pipeline parallelism is a key technique for training large language models in distributed settings, but existing schedules face a fundamental trade off between hardware utilization and optimization consistency. Synchronous methods preserve correct optimization but suffer from pipeline bubbles, while asynchronous methods eliminate these bubbles at the cost of stale model weights and additional mechanisms such as weight stashing or prediction. In this work, we present PACI (Pipeline Asynchronous training with Controlled Inconsistency), an asynchronous pipeline training algorithm that achieves bubble free execution while maintaining bounded weight inconsistency without weight stashing, weight prediction, additional parameter copies, or global synchronization. Evaluated on GPT-2 pretraining, PACI matches the convergence, model quality, and memory footprint of a synchronous schedule while reducing training time-to-accuracy by up to 1.69x.

Many drugs produce clear clinical or cellular effects, but the molecular targets underlying these effects often remain unknown. Identifying these targets is essential for understanding mechanisms of action, anticipating side effects, and guiding drug optimization.

Direct biochemical validation methods can provide strong evidence for drug–target interactions, but they are often time-consuming, expensive, and difficult to apply at large scale. This creates a need for high-throughput approaches that can prioritize candidate target proteins for focused downstream validation. Limited Proteolysis Coupled to Mass Spectrometry, LiP-MS, is a proteome-wide approach for target identification that detects drug-induced changes in protein structure or accessibility through changes in protease-generated peptide patterns.

However, existing LiP-MS scoring strategies have mainly been tested on small sets of well-characterized, high-affinity ligands, leaving their performance on broader and more heterogeneous drug–target panels unclear. In this thesis, we evaluate LiP-MS-based target identification on a diverse panel of drug–target pairs and develop a computational framework for improving target prioritization.

The framework combines per-drug statistical evidence with cross-drug background patterns in order to recover known target proteins while reducing the impact of recurrent false-positive responders. We further incorporate biological knowledge about protein structure and function by testing whether high-scoring peptides fall within known active or binding domains. This additional layer helps distinguish biologically plausible candidates from proteins whose statistical signal is less likely to reflect a true drug–target interaction.

Together, this work provides a systematic analysis of factors affecting LiP-MS target identification and introduces a pipeline that integrates statistical scoring with protein domain information. By improving recovery of known targets and supporting more confident interpretation of ambiguous hits, the proposed framework can help guide future validation experiments toward the most relevant candidate target proteins.

Scaffold decoration is a central task in structure-based drug design: starting from a fixed molecular core, the goal is to add peripheral fragments that improve target binding while preserving drug-like properties. Existing scaffold-completion methods often ignore the three-dimensional structure of the target protein, whereas structure-aware generative models are typically designed for unconstrained de novo generation rather than scaffold-preserving lead elaboration. We introduce MODOLO (Molecular Docking-based Lead Elaboration), an attachment-point-centric generative model for target-aware scaffold decoration. Given an incomplete scaffold, its docking pose, and the surrounding protein pocket, MODOLO represents the protein-ligand complex as a sparse heterogeneous 3D graph and uses grouped vector attention to encode local geometric interactions around each attachment point. A VAE-based Transformer then generates SMILES fragments conditioned on this pocket-aware representation, which are reattached to the scaffold to form complete molecules. On CrossDock2020, MODOLO achieves the best overall performance across affinity, pose score, scaffold similarity, synthetic accessibility, and success rate, improving success from 0.17 for the strongest non-random baseline to 0.41. The model also generalizes to an out-of-distribution PoseBusters subset, supporting its utility for structure-based lead elaboration on unseen targets.

DNA-based storage offers exceptional density and durability, but its scalability depends on efficiently synthesizing many strands in parallel under physical constraints. We study two complementary models capturing limitations arising from nucleotide availability and spatial interactions in large strand arrays.

First, we introduce complex synthesis sequences, a hybrid framework that bridges enzymatic synthesis, which permits strand-specific nucleotide additions, and photolithographic synthesis, which applies common additions across many strands. We characterize the achievable information rate through an analogue of the deletion ball, derive tight expressions for the maximal rate and its asymptotic behavior, and present a dynamic programming algorithm for computing an optimal synthesis sequence for known strands.

We then consider a two-dimensional array model motivated by optical and chemical coupling, in which strands follow a fixed global synthesis sequence and at most one strand per row may advance in each cycle. For two strands in a single row, we show that the laggard-first policy is asymptotically optimal among online policies without look-ahead, while one-symbol look-ahead yields a strict improvement in the binary case.

We further show that even globally optimal scheduling incurs an unavoidable expected overhead that grows linearly with strand length. These results are complemented by an offline dynamic programming algorithm and a constant-redundancy binary coding scheme with a deterministic worst-case synthesis-time guarantee. Together, the two works establish a unified framework for understanding the information-theoretic and scheduling limits of constrained DNA synthesis.

My research focuses on mechanistic interpretability-understanding and improving multi-modal models by analyzing their internal computations. I will present recent studies, including Diffusion Lens (ACL 2024), Padding Tone (NAACL 2025), and Follow the Flow (ACL 2026), and DeLeaker (ICL 2026). These works propose methods to reveal what text-to-image models learn and how computations are performed internally. These insights are then applied to solve concrete problems, such as semantic leakage, paving the way for more reliable and transparent generative AI.