arXiv subjects

Maodong Li

Publications and source records attributed to Maodong Li.

13 recordsLinked to original sources

arXiv2026

Exploiting Intrinsic Duality for Multi-Hop Question Generation

Multi hop question generation (MQG) aims to generate questions from multiple given documents and target answers, whereas question answering (QA) focuses on deriving answers from documents given specific questions. Although MQG and QA are inherently dual tasks, most existing MQG studies largely overlook this intrinsic duality. To address this limitation, we propose QQ, a novel framework that exploits the duality between Question and answer for multi hop Question generation. Specifically, QQ employs a unified architecture functioning simultaneously as both an MQG and a QA model to fully leverage their interdependence. Our framework is driven by two key mechanisms: (i) enforcing bidirectional alignment constraints to ensure strict mutual correspondence between the questions generated by the MQG model and the answers produced by the QA model; and (ii) applying contrastive learning to pull paired question answer representations closer while pushing unpaired ones apart, thereby reinforcing this correspondence. Extensive automatic and human evaluations on the HotpotQA and MuSiQue datasets demonstrate that the QQ framework significantly improves the quality of generated multi hop questions.

cs.CL↗

arXiv2026

Enhancing Target-Guided Proactive Dialogue Systems via Conversational Scenario Modeling and Intent-Keyword Bridging

A target-guided proactive dialogue system aims to steer conversations proactively toward pre-defined targets, such as designated keywords or specific topics. During guided conversations, dynamically modeling conversational scenarios and intent keywords to guide system utterance generation is beneficial; however, existing work largely overlooks this aspect, resulting in a mismatch with the dynamics of real-world conversations. In this paper, we jointly model user profiles and domain knowledge as conversational scenarios to introduce a scenario bias that dynamically influences system utterances, and employ intent-keyword bridging to predict intent keywords for upcoming dialogue turns, providing higher level and more flexible guidance. Extensive automatic and human evaluations demonstrate the effectiveness of conversational scenario modeling and intent keyword bridging, yielding substantial improvements in proactivity, fluency, and informativeness for target-guided proactive dialogue systems, thereby narrowing the gap with real world interactions.

cs.CL↗

arXiv2026

A Massively Scalable Ligand-Protein Dissociation Dynamic Database Derived from Atomistic Molecular Modelling

Understanding the kinetics of drug-protein interactions is paramount for drug design, yet the field lacks large-scale, dynamic data to move beyond static structural analysis. Here, we present DD-03B, a massively scalable database providing dynamic, all-atom dissociation trajectories for a broad set of ligand-protein complexes. Utilising and extending a validated computational pipeline, we generated dissociation trajectories for 19,037 ligand-protein complexes sourced from PDBbind+v2020R1, resulting in a repository of approximately 0.3 billion simulation frames totalling 40 TB in size. For these systems-which possess experimental binding affinities (kd) but typically lack measured koff rates-we computed and assigned dissociation rate constants through trajectory reweighting. Our analysis reveals that protein-ligand complexes can be categorised into three mechanistic types (pathway-dominant, open-pocket, and entropy-pocket systems), each requiring distinct strategies for accurate kinetic characterisation. Together with our previously released DD-13M, DD-03B forms the core of the expandable Dissociation Dynamic Database (DDD) project, which will be continuously augmented with new trajectories. This large-scale, publicly available resource establishes a critical foundation for training and benchmarking next-generation generative AI models to predict and optimise drug-protein dissociation kinetics.

physics.comp-ph↗

arXiv2026

Pseudo-Siamese Network for Planning in Target-Oriented Proactive Dialogues

A target-oriented proactive dialogue system is designed to steer conversations toward predefined targets while actively providing suggestions. The core paradigm of such a system is to plan a reasonable dialogue path and subsequently guide language models (e.g., pre-trained or large language models) to generate responses, where dialogue path planning serves as the central component-a novel yet under-explored problem. In this work, we propose a Forward-Focused Bidirectional Pseudo-Siamese Network (FF-BPSN) for dialogue path planning toward predefined dialogue targets. FF-BPSN employs two identical transformer-based decoders for forward and backward planning, together with a forward-focused module that integrates bidirectional information to construct the final forward path. This path benefits from bidirectional planning while prioritizing forward information. We then employ the planned path to guide language models in response generation. Extensive experiments on DuRecDial and DuRecDial 2.0 demonstrate that FF-BPSN achieves state-of-the-art performance in dialogue path planning and significantly enhances the effectiveness of target-oriented proactive dialogue systems.

cs.CL↗

arXiv2025

Great Restraining Wall in Multidimensional Collective Variable Space

Enhanced sampling methods are pivotal for exploring rare events in molecular dynamics (MD), yet face challenges in high-dimensional collective variable (CV) spaces where exhaustive sampling becomes computationally prohibitive. While techniques like metadynamics (MetaD) and path-CV enable targeted free energy surface (FES) reconstruction, they often struggle with confinement stability, hyperparameter sensitivity, and geometric flexibility. This work introduces the Great Restraining Wall (GW) method, a robust framework for efficient FES sampling within predefined CV subspaces, addressing these limitations through a novel kernel density estimation (KDE)-derived restraining potential. GW operates by constructing a bias potential that confines sampling to user defined regions ranging from multidimensional masks to 1D pathways via asymptotically half-harmonic barriers. Unlike MetaD variants requiring iterative bias deposition, GW potential is derived from a cumulative distribution function, ensuring confinement without manual hyperparameter tuning. GW provides a versatile, stable, and efficient framework for targeted FES sampling, particularly beneficial for complex biomolecular systems with intricate CV landscapes. Its integration with existing enhanced sampling protocols opens avenues for studying ligand binding, conformational transitions, and other rare events with unprecedented precision. Future work will explore GW extension to adaptive regions and machine learning-guided CV discovery.

physics.comp-ph↗

arXiv2025

Multi-Hop Question Generation via Dual-Perspective Keyword Guidance

Multi-hop question generation (MQG) aims to generate questions that require synthesizing multiple information snippets from documents to derive target answers. The primary challenge lies in effectively pinpointing crucial information snippets related to question-answer (QA) pairs, typically relying on keywords. However, existing works fail to fully utilize the guiding potential of keywords and neglect to differentiate the distinct roles of question-specific and document-specific keywords. To address this, we define dual-perspective keywords (i.e., question and document keywords) and propose a Dual-Perspective Keyword-Guided (DPKG) framework, which seamlessly integrates keywords into the multi-hop question generation process. We argue that question keywords capture the questioner's intent, whereas document keywords reflect the content related to the QA pair. Functionally, question and document keywords work together to pinpoint essential information snippets in the document, with question keywords required to appear in the generated question. The DPKG framework consists of an expanded transformer encoder and two answer-aware transformer decoders for keyword and question generation, respectively. Extensive experiments demonstrate the effectiveness of our work, showcasing its promising performance and underscoring its significant value in the MQG task.

cs.CL↗

arXiv2025

A Novel 4-D Dataset Paradigm for Studying Complete Ligand-Protein Dissociation Dynamics

The kinetics and dynamics of drug-protein binding and dissociation are crucial to understanding drug absorption and metabolism. Despite advances in artificial intelligence (AI) tools for drug-protein interaction studies, existing training datasets remain limited to static structures or quasi-static conformations. This paper proposes a novel computational approach for rapidly generating drug-protein dissociation trajectories and presents the inaugural dynamically time-resolved 4-D (t, x, y, z) trajectory database DD-13M. This dataset captures over 26,000 complete dissociation processes for 565 ligand-protein complexes, providing nearly 13 million frames of all-atom simulation trajectories. A deep equivariant generative model, UnbindingFlow, was trained using the DD-13M dataset. This model has the capacity to produce dissociation trajectories for novel targets whilst accurately predicting their rate constants (koff). DD-13M introduces a new type of training dataset for AI models, establishing a de novo paradigm for studying the dynamics of drug-protein interactions.

physics.comp-ph↗

arXiv2025

Performing Path Integral Molecular Dynamics Using Artificial Intelligence Enhanced Molecular Simulation Framework

This study employed an artificial intelligence-enhanced molecular simulation framework to enable efficient Path Integral Molecular Dynamics (PIMD) simulations. Owing to its modular architecture and high-throughput capabilities, the framework effectively mitigates the computational complexity and resource-intensive limitations associated with conventional PIMD approaches. By integrating machine learning force fields (MLFFs) into the framework, we rigorously tested its performance through two representative cases: a small-molecule reaction system (double proton transfer in formic acid dimer) and a bulk-phase transition system (water-ice phase transformation). Computational results demonstrate that the proposed framework achieves accelerated PIMD simulations while preserving quantum mechanical accuracy. These findings show that nuclear quantum effects can be captured for complex molecular systems, using relatively low computational cost.

physics.chem-ph↗

arXiv2024

A Sinking Approach to Explore Arbitrary Areas in Free Energy Landscapes

To address the time-scale limitations in molecular dynamics (MD) simulations, numerous enhanced sampling methods have been developed to expedite the exploration of complex free energy landscapes. A commonly employed approach accelerates the sampling of degrees of freedom associated with pre-defined collective variables (CVs), which typically tends to traverse the entire CV range. However, in many scenarios, the focus of interest is on specific regions within the CV space. This paper introduces a novel "sinking" approach that enables enhanced sampling of arbitrary areas within the CV space. We begin by proposing a gridded convolutional approximation that productively replicates the effects of metadynamics, a powerful CV-based enhanced sampling technique. Building on this, we present the SinkMeta method, which "sinks" the interior bias potential to create restraining potential "cliffs" at the grid edges. This technique can confine the exploration of CVs in MD simulations to a preset area. Our experimental results demonstrate that SinkMeta requires minimal sampling steps to estimate the free energy landscape for CV subspaces of various shapes and dimensions, including irregular two-dimensional regions and one-dimensional pathways between metastable states. We believe that SinkMeta will pioneer a new paradigm for sampling partial phase spaces, especially offering an efficient and flexible solution for sampling minimum free energy paths in high-dimensional spaces.

physics.comp-ph↗

arXiv2023

A Generalized Nucleation Theory for Ice Crystallization

Despite the simplicity of the water molecule, the kinetics of ice nucleation under natural conditions can be complex. We investigated spontaneously grown ice nuclei using all-atom molecular dynamics simulations and found significant differences between the kinetics of ice formation through spontaneously formed and ideal nuclei. Since classical nucleation theory can only provide a good description of ice nucleation in ideal conditions, we propose a generalized nucleation theory that can better characterize the kinetics of ice crystal nucleation in general conditions. This study provides an explanation on why previous experimental and computational studies have yielded widely varying critical nucleation sizes.

cond-mat.soft↗

arXiv2021

Atomistic View of Homogeneous Nucleation of Water into Polymorphic Ices

Water is one of the most abundant substances on Earth, and ice, i.e., solid water, has more than 18 known phases. Normally ice in nature exists only as Ice Ih, Ice Ic, or a stacking disordered mixture of both. Although many theoretical efforts have been devoted to understanding the thermodynamics of different ice phases at ambient temperature and pressure, there still remains many puzzles. We simulated the reversible transitions between water and different ice phases by performing full atom molecular dynamics simulations. Using the enhanced sampling method MetaITS with the two selected X-ray diffraction peak intensities as collective variables, the ternary phase diagrams of liquid water, ice Ih, ice Ic at multiple were obtained. We also present a simple physical model which successfully explains the thermodynamic stability of ice. Our results agree with experiments and leads to a deeper understanding of the ice nucleation mechanism.

cond-mat.stat-mech↗

arXiv2020

Deep Reinforcement Learning of Transition States

Combining reinforcement learning (RL) and molecular dynamics (MD) simulations, we propose a machine-learning approach (RL$^‡$) to automatically unravel chemical reaction mechanisms. In RL$^‡$, locating the transition state of a chemical reaction is formulated as a game, where a virtual player is trained to shoot simulation trajectories connecting the reactant and product. The player utilizes two functions, one for value estimation and the other for policy making, to iteratively improve the chance of winning this game. We can directly interpret the reaction mechanism according to the value function. Meanwhile, the policy function enables efficient sampling of the transition paths, which can be further used to analyze the reaction dynamics and kinetics. Through multiple experiments, we show that RL‡ can be trained tabula rasa hence allows us to reveal chemical reaction mechanisms with minimal subjective biases.

physics.chem-ph↗

arXiv2020

A Perspective on Deep Learning for Molecular Modeling and Simulations

Deep learning is transforming many areas in science, and it has great potential in modeling molecular systems. However, unlike the mature deployment of deep learning in computer vision and natural language processing, its development in molecular modeling and simulations is still at an early stage, largely because the inductive biases of molecules are completely different from those of images or texts. Footed on these differences, we first reviewed the limitations of traditional deep learning models from the perspective of molecular physics, and wrapped up some relevant technical advancement at the interface between molecular modeling and deep learning. We do not focus merely on the ever more complex neural network models, instead, we emphasize the theories and ideas behind modern deep learning. We hope that transacting these ideas into molecular modeling will create new opportunities. For this purpose, we summarized several representative applications, ranging from supervised to unsupervised and reinforcement learning, and discussed their connections with the emerging trends in deep learning. Finally, we outlook promising directions which may help address the existing issues in the current framework of deep molecular modeling.

physics.comp-ph↗