arXiv subjects

Nirav Pravinbhai Bhatt

Publications and source records attributed to Nirav Pravinbhai Bhatt.

7 recordsLinked to original sources

arXiv2026

A Large Scale Investigation of Scaling Limits in Chemical Language Models

Chemical Language Models (CLMs) are increasingly used in de novo drug design, driven by recent growth in model scale, compute, and dataset size. However, the relationship between design choices, training dynamics, and downstream generation quality remains poorly understood. We present a compute-controlled scaling study of CLMs comprising more than 30,000 experiments across molecular representations (SMILES, SELFIES, SAFE), tokenizations (atom-level and byte-pair encoding), model scales (0.5M-1B parameters), leakage-controlled datasets (MOSES, ChEMBL, PubChem, ZINC-22), and architectures (decoder-only and encoder-decoder). By fitting IsoFLOP profiles, we establish clear scaling trends in pretraining loss, but find that these improvements do not translate into comparable gains in goal-directed molecular design. Layer-wise probing and sparse autoencoder analysis reveal continued development of chemical representations: chemical syntax saturates early, while semantic properties emerge more slowly and become increasingly accessible with further training and model scale. These representational gains coexist with diminishing improvements in goal-directed generation under the evaluated protocols and oracle budgets. Our resulting suite of models, NovoMolGen, achieves state-of-the-art results, outperforming prior CLMs and specialized generative models in goal-directed molecular generation across drug discovery tasks. These findings expose a disconnect between chemical representation learning and downstream molecular design, motivating the development of pretraining paradigms that more directly learn chemical semantics.

cs.LG↗

arXiv2026

An Integrative Computational Approach to Predict Viral Epitopes by Targeting the MHC-TCR Complexation

T-cell immunity acts as a major defense system against controlling viral infections in vertebrates. During viral entry, innate immune cells degrade the viral proteins (antigens) and present them on their surface via Major Histocompatibility (MHC) proteins. T-cell receptors (TCRs) recognize these antigens/peptides presented by MHC (pMHC), initiating a T-cell mediated immune response. Despite its significance, the mechanism by which pMHC-TCR binding triggers T-cell activation remains unclear. In this study, we employed an integrative computational approach combining Bioinformatics, Molecular Dynamics (MD) simulations, and Machine Learning (ML) to identify viral epitopes as potential vaccine candidates. We performed large-scale all-atom and coarse-grained MD simulations on MHC-peptide-TCR complexes embedded into dendritic and T-cells, for which experimental immunogenicity data is available. One hundred fifty such systems are simulated for 1 μs each to capture the conformational and dynamical changes that underlie T-cell activation. Our ML model (DynamiT), trained on simulation-derived structural and dynamical features extracted from 2500 time points, revealed key determinants responsible for T-cell activation with an accuracy of 73.3%. Notably, we have identified the bending of the TCR transmembrane region, major dynamic motions of the TCRα constant region and the buried surface area at the pMHC and TCR interface as critical factors influencing immune response initiation. Our approach unravels the mechanism of T-cell mediated immune response and helps ML-guided screening of viral epitopes for vaccine development.

q-bio.BM↗

arXiv2026

OSIL: Learning Offline Safe Imitation Policies with Safety Inferred from Non-preferred Trajectories

This work addresses the problem of offline safe imitation learning (IL), where the goal is to learn safe and reward-maximizing policies from demonstrations that do not have per-timestep safety cost or reward information. In many real-world domains, online learning in the environment can be risky, and specifying accurate safety costs can be difficult. However, it is often feasible to collect trajectories that reflect undesirable or unsafe behavior, implicitly conveying what the agent should avoid. We refer to these as non-preferred trajectories. We propose a novel offline safe IL algorithm, OSIL, that infers safety from non-preferred demonstrations. We formulate safe policy learning as a Constrained Markov Decision Process (CMDP). Instead of relying on explicit safety cost and reward annotations, OSIL reformulates the CMDP problem by deriving a lower bound on reward maximizing objective and learning a cost model that estimates the likelihood of non-preferred behavior. Our approach allows agents to learn safe and reward-maximizing behavior entirely from offline demonstrations. We empirically demonstrate that our approach can learn safer policies that satisfy cost constraints without degrading the reward performance, thus outperforming several baselines.

cs.LG↗

arXiv2025

BiCA: Effective Biomedical Dense Retrieval with Citation-Aware Hard Negatives

Hard negatives are essential for training effective retrieval models. Hard-negative mining typically relies on ranking documents using cross-encoders or static embedding models based on similarity metrics such as cosine distance. Hard negative mining becomes challenging for biomedical and scientific domains due to the difficulty in distinguishing between source and hard negative documents. However, referenced documents naturally share contextual relevance with the source document but are not duplicates, making them well-suited as hard negatives. In this work, we propose BiCA: Biomedical Dense Retrieval with Citation-Aware Hard Negatives, an approach for hard-negative mining by utilizing citation links in 20,000 PubMed articles for improving a domain-specific small dense retriever. We fine-tune the GTE_small and GTE_Base models using these citation-informed negatives and observe consistent improvements in zero-shot dense retrieval using nDCG@10 for both in-domain and out-of-domain tasks on BEIR and outperform baselines on long-tailed topics in LoTTE using Success@5. Our findings highlight the potential of leveraging document link structure to generate highly informative negatives, enabling state-of-the-art performance with minimal fine-tuning and demonstrating a path towards highly data-efficient domain adaptation.

cs.IR↗

arXiv2025

SafeMIL: Learning Offline Safe Imitation Policy from Non-Preferred Trajectories

In this work, we study the problem of offline safe imitation learning (IL). In many real-world settings, online interactions can be risky, and accurately specifying the reward and the safety cost information at each timestep can be difficult. However, it is often feasible to collect trajectories reflecting undesirable or risky behavior, implicitly conveying the behavior the agent should avoid. We refer to these trajectories as non-preferred trajectories. Unlike standard IL, which aims to mimic demonstrations, our agent must also learn to avoid risky behavior using non-preferred trajectories. In this paper, we propose a novel approach, SafeMIL, to learn a parameterized cost that predicts if the state-action pair is risky via Multiple Instance Learning. The learned cost is then used to avoid non-preferred behaviors, resulting in a policy that prioritizes safety. We empirically demonstrate that our approach can learn a safer policy that satisfies cost constraints without degrading the reward performance, thereby outperforming several baselines.

cs.LG↗

arXiv2025

Learning from Observation: A Survey of Recent Advances

Imitation Learning (IL) algorithms offer an efficient way to train an agent by mimicking an expert's behavior without requiring a reward function. IL algorithms often necessitate access to state and action information from expert demonstrations. Although expert actions can provide detailed guidance, requiring such action information may prove impractical for real-world applications where expert actions are difficult to obtain. To address this limitation, the concept of learning from observation (LfO) or state-only imitation learning (SOIL) has recently gained attention, wherein the imitator only has access to expert state visitation information. In this paper, we present a framework for LfO and use it to survey and classify existing LfO methods in terms of their trajectory construction, assumptions and algorithm's design choices. This survey also draws connections between several related fields like offline RL, model-based RL and hierarchical RL. Finally, we use our framework to identify open problems and suggest future research directions.

cs.LG↗

arXiv2025

Functional Groups are All you Need for Chemically Interpretable Molecular Property Prediction

Molecular property prediction using deep learning (DL) models has accelerated drug and materials discovery, but the resulting DL models often lack interpretability, hindering their adoption by chemists. This work proposes developing molecule representations using the concept of Functional Groups (FG) in chemistry. We introduce the Functional Group Representation (FGR) framework, a novel approach to encoding molecules based on their fundamental chemical substructures. Our method integrates two types of functional groups: those curated from established chemical knowledge (FG), and those mined from a large molecular corpus using sequential pattern mining (MFG). The resulting FGR framework encodes molecules into a lower-dimensional latent space by leveraging pre-training on a large dataset of unlabeled molecules. Furthermore, the proposed framework allows the inclusion of 2D structure-based descriptors of molecules. We demonstrate that the FGR framework achieves state-of-the-art performance on a diverse range of 33 benchmark datasets spanning physical chemistry, biophysics, quantum mechanics, biological activity, and pharmacokinetics while enabling chemical interpretability. Crucially, the model's representations are intrinsically aligned with established chemical principles, allowing chemists to directly link predicted properties to specific functional groups and facilitating novel insights into structure-property relationships. Our work presents a significant step toward developing high-performing, chemically interpretable DL models for molecular discovery.

cs.LG↗