arXiv subjects

Junmei Wang

Publications and source records attributed to Junmei Wang.

10 recordsLinked to original sources

arXiv2026

Localize, Then Reason: Visual Latent Structural Reasoning for Molecular Properties and Edits

Local chemical perception and property reasoning are both essential for understanding how molecular structure determines properties. Current LLM-based chemical reasoning methods either receive SMILES/molecular images together with descriptions of local motifs, or reason directly from molecular images. Neither approach enables the model to focus on chemically meaningful regions before reasoning. To address this gap, we propose Visual Latent Structural Reasoning (VLSR), an end-to-end framework that jointly learns localization and reasoning from molecular images. Central to our approach is a localize-then-reason strategy. VLSR first learns to locate chemically meaningful regions in a molecular image. It then reasons about their property effects in a compact latent workspace before producing the final answer. Under the same inference setup, this design achieves 9.6X higher throughput than a comparable textual-reasoning baseline.

cs.CL↗

arXiv2026

Learning Mechanistic Reasoning for Chemical Reactions with Large Language Models

Reaction mechanisms consist of the step-by-step sequences of elementary reactions that explain chemical transformations. Learning the mechanism logic is therefore essential for enhancing the fundamental chemical intelligence of large language models (LLMs). The stepwise deduction of reaction mechanism aligns naturally with the reasoning paradigms of reasoning LLMs. However, current chemical LLMs primarily emphasize coarse-grained name reactions for product prediction and retrosynthesis, often leading to physical inconsistencies and hallucinations. In contrast, specialized small-scale generative models for mechanism inference typically suffer from restricted generalization capacity across diverse chemical spaces. To overcome these limitations, we built a novel, large-scale reasoning dataset of reaction mechanisms. Furthermore, we established the FukuyamaBench, a difficult benchmark derived from Fukuyama's Advanced Organic Reaction Mechanism book, to rigorously evaluate model performance on hierarchical mechanism reasoning. Our fine-tuned Qwen3-30B-A3B achieves 8.3% exact pathway match on FukuyamaBench Set~A, surpassing the specialized FlowER model (5.1%), demonstrating that mechanism-aware training substantially enhances chemical reasoning in language models.

cs.LG↗

arXiv2026

Fine-tuning DeepSeek-OCR-2 for Molecular Structure Recognition

Optical Chemical Structure Recognition (OCSR) is critical for converting 2D molecular diagrams from printed literature into machine-readable formats. While Vision-Language Models have shown promise in end-to-end OCR tasks, their direct application to OCSR remains challenging, and direct full-parameter supervised fine-tuning often fails. In this work, we adapt DeepSeek-OCR-2 for molecular optical recognition by formulating the task as image-conditioned SMILES generation. To overcome training instabilities, we propose a two-stage progressive supervised fine-tuning strategy: starting with parameter-efficient LoRA and transitioning to selective full-parameter fine-tuning with split learning rates. We train our model on a large-scale corpus combining synthetic renderings from PubChem and realistic patent images from USPTO-MOL to improve coverage and robustness. Our fine-tuned model, MolSeek-OCR, demonstrates competitive capabilities, achieving exact matching accuracies comparable to the best-performing image-to-sequence model. However, it remains inferior to state-of-the-art image-to-graph modelS. Furthermore, we explore reinforcement-style post-training and data-curation-based refinement, finding that they fail to improve the strict sequence-level fidelity required for exact SMILES matching.

cs.CV↗

arXiv2026

Ultrasound-CLIP: Semantic-Aware Contrastive Pre-training for Ultrasound Image-Text Understanding

Ultrasound imaging is widely used in clinical diagnostics due to its real-time capability and radiation-free nature. However, existing vision-language pre-training models, such as CLIP, are primarily designed for other modalities, and are difficult to directly apply to ultrasound data, which exhibit heterogeneous anatomical structures and diverse diagnostic attributes. To bridge this gap, we construct US-365K, a large-scale ultrasound image-text dataset containing 365k paired samples across 52 anatomical categories. We establish Ultrasonographic Diagnostic Taxonomy (UDT) containing two hierarchical knowledge frameworks. Ultrasonographic Hierarchical Anatomical Taxonomy standardizes anatomical organization, and Ultrasonographic Diagnostic Attribute Framework formalizes nine diagnostic dimensions, including body system, organ, diagnosis, shape, margins, echogenicity, internal characteristics, posterior acoustic phenomena, and vascularity. Building upon these foundations, we propose Ultrasound-CLIP, a semantic-aware contrastive learning framework that introduces semantic soft labels and semantic loss to refine sample discrimination. Moreover, we construct a heterogeneous graph modality derived from UDAF's textual representations, enabling structured reasoning over lesion-attribute relations. Extensive experiments with patient-level data splitting demonstrate that our approach achieves state-of-the-art performance on classification and retrieval benchmarks, while also delivering strong generalization to zero-shot, linear probing, and fine-tuning tasks.

cs.CV↗

arXiv2025

GED-Consistent Disentanglement of Aligned and Unaligned Substructures for Graph Similarity Learning

Graph Similarity Computation (GSC) is a fundamental graph related task where Graph Edit Distance (GED) serves as a prevalent metric. GED is determined by an optimal alignment between a pair of graphs that partitions each into aligned (zero-cost) and unaligned (cost-incurring) substructures. Due to NP-hard nature of exact GED computation, GED approximations based on Graph Neural Network(GNN) have emerged. Existing GNN-based GED approaches typically learn node embeddings for each graph and then aggregate pairwise node similarities to estimate the final similarity. Despite their effectiveness, we identify a mismatch between this prevalent node-centric matching paradigm and the core principles of GED. This discrepancy leads to two critical limitations: (1) a failure to capture the global structural correspondence for optimal alignment, and (2) a misattribution of edit costs driven by spurious node level signals. To address these limitations, we propose GCGSim, a GED-consistent graph similarity learning framework centering on graph-level matching and substructure-level edit costs. Specifically, we make three core technical contributions. Extensive experiments on four benchmark datasets show that GCGSim achieves state-of-the-art performance. Our comprehensive analyses further validate that the framework effectively learns disentangled and semantically meaningful substructure representations.

cs.LG↗

arXiv2025

Auxiliary Discrminator Sequence Generative Adversarial Networks (ADSeqGAN) for Few Sample Molecule Generation

In this work, we introduce Auxiliary Discriminator Sequence Generative Adversarial Networks (ADSeqGAN), a novel approach for molecular generation in small-sample datasets. Traditional generative models often struggle with limited training data, particularly in drug discovery, where molecular datasets for specific therapeutic targets, such as nucleic acids binders and central nervous system (CNS) drugs, are scarce. ADSeqGAN addresses this challenge by integrating an auxiliary random forest classifier as an additional discriminator into the GAN framework, significantly improves molecular generation quality and class specificity. Our method incorporates pretrained generator and Wasserstein distance to enhance training stability and diversity. We evaluate ADSeqGAN across three representative cases. First, on nucleic acid- and protein-targeting molecules, ADSeqGAN shows superior capability in generating nucleic acid binders compared to baseline models. Second, through oversampling, it markedly improves CNS drug generation, achieving higher yields than traditional de novo models. Third, in cannabinoid receptor type 1 (CB1) ligand design, ADSeqGAN generates novel druglike molecules, with 32.8\% predicted actives surpassing hit rates of CB1-focused and general-purpose libraries when assessed by a target-specific LRIP-SF scoring function. Overall, ADSeqGAN offers a versatile framework for molecular design in data-scarce scenarios, with demonstrated applications in nucleic acid binders, CNS drugs, and CB1 ligands.

cs.LG↗

arXiv2024

Utilizing BERT for Information Retrieval: Survey, Applications, Resources, and Challenges

Recent years have witnessed a substantial increase in the use of deep learning to solve various natural language processing (NLP) problems. Early deep learning models were constrained by their sequential or unidirectional nature, such that they struggled to capture the contextual relationships across text inputs. The introduction of bidirectional encoder representations from transformers (BERT) leads to a robust encoder for the transformer model that can understand the broader context and deliver state-of-the-art performance across various NLP tasks. This has inspired researchers and practitioners to apply BERT to practical problems, such as information retrieval (IR). A survey that focuses on a comprehensive analysis of prevalent approaches that apply pretrained transformer encoders like BERT to IR can thus be useful for academia and the industry. In light of this, we revisit a variety of BERT-based methods in this survey, cover a wide range of techniques of IR, and group them into six high-level categories: (i) handling long documents, (ii) integrating semantic information, (iii) balancing effectiveness and efficiency, (iv) predicting the weights of terms, (v) query expansion, and (vi) document expansion. We also provide links to resources, including datasets and toolkits, for BERT-based IR systems. A key highlight of our survey is the comparison between BERT's encoder-based models and the latest generative Large Language Models (LLMs), such as ChatGPT, which rely on decoders. Despite the popularity of LLMs, we find that for specific tasks, finely tuned BERT encoders still outperform, and at a lower deployment cost. Finally, we summarize the comprehensive outcomes of the survey and suggest directions for future research in the area.

cs.IR↗

arXiv2024

Emerging Opportunities of Using Large Language Models for Translation Between Drug Molecules and Indications

A drug molecule is a substance that changes the organism's mental or physical state. Every approved drug has an indication, which refers to the therapeutic use of that drug for treating a particular medical condition. While the Large Language Model (LLM), a generative Artificial Intelligence (AI) technique, has recently demonstrated effectiveness in translating between molecules and their textual descriptions, there remains a gap in research regarding their application in facilitating the translation between drug molecules and indications, or vice versa, which could greatly benefit the drug discovery process. The capability of generating a drug from a given indication would allow for the discovery of drugs targeting specific diseases or targets and ultimately provide patients with better treatments. In this paper, we first propose a new task, which is the translation between drug molecules and corresponding indications, and then test existing LLMs on this new task. Specifically, we consider nine variations of the T5 LLM and evaluate them on two public datasets obtained from ChEMBL and DrugBank. Our experiments show the early results of using LLMs for this task and provide a perspective on the state-of-the-art. We also emphasize the current limitations and discuss future work that has the potential to improve the performance on this task. The creation of molecules from indications, or vice versa, will allow for more efficient targeting of diseases and significantly reduce the cost of drug discovery, with the potential to revolutionize the field of drug discovery in the era of generative AI.

cs.AI↗

arXiv2020

Efficient Formulation of Polarizable Gaussian Multipole Electrostatics for Biomolecular Simulations

Molecular dynamics simulations of biomolecules have been widely adopted in biomedical studies. As classical point-charge models continue to be used in routine biomolecular applications, there have been growing demands on developing polarizable force fields for handling more complicated biomolecular processes. Here we focus on a recently proposed polarizable Gaussian Multipole (pGM) model for biomolecular simulations. A key benefit of pGM is its screening of all short-range electrostatic interactions in a physically consistent manner, which is critical for stable charge-fitting and is needed to reproduce molecular anisotropy. Another advantage of pGM is that each atom's multipoles are represented by a single Gaussian function or its derivatives, allowing for more efficient electrostatics than other Gaussian-based models. In this study we present an efficient formulation for the pGM model defined with respect to a local frame formed with a set of covalent basis vectors. The covalent basis vectors are chosen to be along each atom's covalent bonding directions. The new local frame allows molecular flexibility during molecular simulations and facilitates an efficient formulation of analytical electrostatic forces without explicit torque computation. Subsequent numerical tests show that analytical atomic forces agree excellently with numerical finite-difference forces for the tested system. Finally, the new pGM electrostatics algorithm is interfaced with the PME implementation in Amber for molecular simulations under the periodic boundary conditions. To validate the overall pGM/PME electrostatics, we conducted an NVE simulation for a small water box of 512 water molecules. Our results show that, to achieve energy conservation in the polarizable model, it is important to ensure enough accuracy on both PME and induction iteration.

physics.comp-ph↗

arXiv2017

A Continuum Poisson-Boltzmann Model for Membrane Channel Proteins

Membrane proteins constitute a large portion of the human proteome and perform a variety of important functions as membrane receptors, transport proteins, enzymes, signaling proteins, and more. The computational studies of membrane proteins are usually much more complicated than those of globular proteins. Here we propose a new continuum model for Poisson-Boltzmann calculations of membrane channel proteins. Major improvements over the existing continuum slab model are as follows: 1) The location and thickness of the slab model are fine-tuned based on explicit-solvent MD simulations. 2) The highly different accessibility in the membrane and water regions are addressed with a two-step, two-probe grid labeling procedure, and 3) The water pores/channels are automatically identified. The new continuum membrane model is optimized (by adjusting the membrane probe, as well as the slab thickness and center) to best reproduce the distributions of buried water molecules in the membrane region as sampled in explicit water simulations. Our optimization also shows that the widely adopted water probe of 1.4 Å for globular proteins is a very reasonable default value for membrane protein simulations. It gives an overall minimum number of inconsistencies between the continuum and explicit representations of water distributions in membrane channel proteins, at least in the water accessible pore/channel regions that we focus on. Finally, we validate the new membrane model by carrying out binding affinity calculations for a potassium channel, and we observe a good agreement with experiment results.

physics.bio-ph↗