arXiv subjects

Zhirong Liu

Publications and source records attributed to Zhirong Liu.

At least 19 recordsLinked to original sources

arXiv2025

AIonopedia: an LLM agent orchestrating multimodal learning for ionic liquid discovery

The discovery of novel Ionic Liquids (ILs) is hindered by critical challenges in property prediction, including limited data, poor model accuracy, and fragmented workflows. Leveraging the power of Large Language Models (LLMs), we introduce AIonopedia, to the best of our knowledge, the first LLM agent for IL discovery. Powered by an LLM-augmented multimodal domain foundation model for ILs, AIonopedia enables accurate property predictions and incorporates a hierarchical search architecture for molecular screening and design. Trained and evaluated on a newly curated and comprehensive IL dataset, our model delivers superior performance. Complementing these results, evaluations on literature-reported systems indicate that the agent can perform effective IL modification. Moving beyond offline tests, the practical efficacy was further confirmed through real-world wet-lab validation, in which the agent demonstrated exceptional generalization capabilities on challenging out-of-distribution tasks, underscoring its ability to accelerate real-world IL discovery.

cs.AI↗

arXiv2024

Efficient and Precise Force Field Optimization for Biomolecules Using DPA-2

Molecular simulations are essential tools in computational chemistry, enabling the prediction and understanding of molecular interactions and thermodynamic properties of biomolecules. However, traditional force fields face significant challenges in accurately representing novel molecules and complex chemical environments due to the labor-intensive process of manually setting optimization parameters and the high computational cost of quantum mechanical calculations. To overcome these difficulties, we fine-tuned a high-accuracy DPA-2 pre-trained model and applied it to optimize force field parameters on-the-fly, significantly reducing computational costs. Our method combines this fine-tuned DPA-2 model with a node-embedding-based similarity metric, allowing seamless augmentation to new chemical species without manual intervention. We applied this process to the TYK2 inhibitor and PTP1B systems and demonstrated its effectiveness through the improvement of free energy perturbation calculation results. This advancement contributes valuable insights and tools for the computational chemistry community.

physics.chem-ph↗

arXiv2024

SE3Set: Harnessing equivariant hypergraph neural networks for molecular representation learning

In this paper, we develop SE3Set, an SE(3) equivariant hypergraph neural network architecture tailored for advanced molecular representation learning. Hypergraphs are not merely an extension of traditional graphs; they are pivotal for modeling high-order relationships, a capability that conventional equivariant graph-based methods lack due to their inherent limitations in representing intricate many-body interactions. To achieve this, we first construct hypergraphs via proposing a new fragmentation method that considers both chemical and three-dimensional spatial information of molecular system. We then design SE3Set, which incorporates equivariance into the hypergragh neural network. This ensures that the learned molecular representations are invariant to spatial transformations, thereby providing robustness essential for accurate prediction of molecular properties. SE3Set has shown performance on par with state-of-the-art (SOTA) models for small molecule datasets like QM9 and MD17. It excels on the MD22 dataset, achieving a notable improvement of approximately 20% in accuracy across all molecules, which highlights the prevalence of complex many-body interactions in larger molecules. This exceptional performance of SE3Set across diverse molecular structures underscores its transformative potential in computational chemistry, offering a route to more accurate and physically nuanced modeling.

cs.LG↗

arXiv2023

Accelerating Kinetics with Time-reversal Path Sampling

In comparison to numerous enhanced sampling methods for equilbrium thermodynamics, accelerating simulations for kinetics and nonequilibrium statistics are relatively rare and less effective. Here we derive a time-reversal path sampling (tRPS) method based on the time reversibility to accelerate simulations for determining the transition rates between free-ernegy basins. It converts the difficult up-hill path sampling into an easy down-hill problem. The method is easy to implement, i.e., forward and backward shooting simulations with opposite initial velocities are conducted from random initial conformations within a transition-state region until they reach the basin minima, which are then assembled to give the distribution of transition paths efficiently. The effects of tRPS are demonstrated by comparison with direct simulations on protein folding and unfolding, where tRPS is shown to give results consistent with direct simulations and increase the efficiency by up to five orders of magnitude. The approach is generally applicable to stochatic processes with microscopic reversibility, no mater whether the variables are continuous or discrete.

physics.chem-ph↗

arXiv2022

Control of phase ordering and elastic properties in phase field crystals through three-point direct correlation

Effects of three-point direct correlation on properties of the phase field crystal (PFC) modeling are examined, for the control of various ordered and disordered phases and their coexistence in both three-dimensional and two-dimensional systems. Such effects are manifested via the corresponding gradient nonlinearity in the PFC free energy functional that is derived from classical density functional theory. Their significant impacts on the stability regimes of ordered phases, phase diagrams, and elastic properties of the system, as compared to those of the original PFC model, are revealed through systematic analyses and simulations. The nontrivial contribution from three-point direct correlation leads to the variation of the critical point of order-disorder transition to which all the phase boundaries in the temperature-density phase diagram converge. It also enables the variation and control of system elastic constants over a substantial range as needed in modeling different types of materials with the same crystalline structure but different elastic properties. The capability of this PFC approach in modeling both solid and soft matter systems is further demonstrated through the effect of three-point correlation on controlling the vapor-liquid-solid coexistence and transitions for body-centered cubic (bcc) phase and on achieving the liquid-stripe or liquid-lamellar phase coexistence. All these provide a valuable and efficient method for the study of structural ordering and evolution in various types of material systems.

cond-mat.mtrl-sci↗

arXiv2021

Reversible Upper Confidence Bound Algorithm to Generate Diverse Optimized Candidates

Most algorithms for the multi-armed bandit problem in reinforcement learning aimed to maximize the expected reward, which are thus useful in searching the optimized candidate with the highest reward (function value) for diverse applications (e.g., AlphaGo). However, in some typical application scenaios such as drug discovery, the aim is to search a diverse set of candidates with high reward. Here we propose a reversible upper confidence bound (rUCB) algorithm for such a purpose, and demonstrate its application in virtual screening upon intrinsically disordered proteins (IDPs). It is shown that rUCB greatly reduces the query times while achieving both high accuracy and low performance loss.The rUCB may have potential application in multipoint optimization and other reinforcement-learning cases.

cs.LG↗

arXiv2021

Retrieval & Interaction Machine for Tabular Data Prediction

Prediction over tabular data is an essential task in many data science applications such as recommender systems, online advertising, medical treatment, etc. Tabular data is structured into rows and columns, with each row as a data sample and each column as a feature attribute. Both the columns and rows of the tabular data carry useful patterns that could improve the model prediction performance. However, most existing models focus on the cross-column patterns yet overlook the cross-row patterns as they deal with single samples independently. In this work, we propose a general learning framework named Retrieval & Interaction Machine (RIM) that fully exploits both cross-row and cross-column patterns among tabular data. Specifically, RIM first leverages search engine techniques to efficiently retrieve useful rows of the table to assist the label prediction of the target row, then uses feature interaction networks to capture the cross-column patterns among the target row and the retrieved rows so as to make the final label prediction. We conduct extensive experiments on 11 datasets of three important tasks, i.e., CTR prediction (classification), top-n recommendation (ranking) and rating prediction (regression). Experimental results show that RIM achieves significant improvements over the state-of-the-art and various baselines, demonstrating the superiority and efficacy of RIM.

cs.IR↗

arXiv2021

AutoFT: Automatic Fine-Tune for Parameters Transfer Learning in Click-Through Rate Prediction

Recommender systems are often asked to serve multiple recommendation scenarios or domains. Fine-tuning a pre-trained CTR model from source domains and adapting it to a target domain allows knowledge transferring. However, optimizing all the parameters of the pre-trained network may result in over-fitting if the target dataset is small and the number of parameters is large. This leads us to think of directly reusing parameters in the pre-trained model which represent more general features learned from multiple domains. However, the design of freezing or fine-tuning layers of parameters requires much manual effort since the decision highly depends on the pre-trained model and target instances. In this work, we propose an end-to-end transfer learning framework, called Automatic Fine-Tuning (AutoFT), for CTR prediction. AutoFT consists of a field-wise transfer policy and a layer-wise transfer policy. The field-wise transfer policy decides how the pre-trained embedding representations are frozen or fine-tuned based on the given instance from the target domain. The layer-wise transfer policy decides how the high?order feature representations are transferred layer by layer. Extensive experiments on two public benchmark datasets and one private industrial dataset demonstrate that AutoFT can significantly improve the performance of CTR prediction compared with state-of-the-art transferring approaches.

cs.IR↗

arXiv2021

Dual Graph enhanced Embedding Neural Network for CTR Prediction

CTR prediction, which aims to estimate the probability that a user will click an item, plays a crucial role in online advertising and recommender system. Feature interaction modeling based and user interest mining based methods are the two kinds of most popular techniques that have been extensively explored for many years and have made great progress for CTR prediction. However, (1) feature interaction based methods which rely heavily on the co-occurrence of different features, may suffer from the feature sparsity problem (i.e., many features appear few times); (2) user interest mining based methods which need rich user behaviors to obtain user's diverse interests, are easy to encounter the behavior sparsity problem (i.e., many users have very short behavior sequences). To solve these problems, we propose a novel module named Dual Graph enhanced Embedding, which is compatible with various CTR prediction models to alleviate these two problems. We further propose a Dual Graph enhanced Embedding Neural Network (DG-ENN) for CTR prediction. Dual Graph enhanced Embedding exploits the strengths of graph representation with two carefully designed learning strategies (divide-and-conquer, curriculum-learning-inspired organized learning) to refine the embedding. We conduct comprehensive experiments on three real-world industrial datasets. The experimental results show that our proposed DG-ENN significantly outperforms state-of-the-art CTR prediction models. Moreover, when applying to state-of-the-art CTR prediction models, Dual graph enhanced embedding always obtains better performance. Further case studies prove that our proposed dual graph enhanced embedding could alleviate the feature sparsity and behavior sparsity problems. Our framework will be open-source based on MindSpore in the near future.

cs.IR↗

arXiv2020

Minimal phase-field crystal modeling of vapor-liquid-solid coexistence and transitions

A phase-field crystal model based on the density-field approach incorporating high-order interparticle direct correlations is developed to study vapor-liquid-solid coexistence and transitions within a single continuum description. Conditions for the realization of the phase coexistence and transition sequence are systematically analyzed and shown to be satisfied by a broad range of model parameters, demonstrating the high flexibility and applicability of the model. Both temperature-density and temperature-pressure phase diagrams are identified, while structural evolution and coexistence among the three phases are examined through dynamical simulations. The model is also able to produce some temperature and pressure related material properties, including effects of thermal expansion and pressure on equilibrium lattice spacing, and temperature dependence of saturation vapor pressure. This model can be used as an effective approach for investigating a variety of material growth and deposition processes based on vapor-solid, liquid-solid, and vapor-liquid-solid growth.

cond-mat.mtrl-sci↗

arXiv2020

A Practical Incremental Method to Train Deep CTR Models

Deep learning models in recommender systems are usually trained in the batch mode, namely iteratively trained on a fixed-size window of training data. Such batch mode training of deep learning models suffers from low training efficiency, which may lead to performance degradation when the model is not produced on time. To tackle this issue, incremental learning is proposed and has received much attention recently. Incremental learning has great potential in recommender systems, as two consecutive window of training data overlap most of the volume. It aims to update the model incrementally with only the newly incoming samples from the timestamp when the model is updated last time, which is much more efficient than the batch mode training. However, most of the incremental learning methods focus on the research area of image recognition where new tasks or classes are learned over time. In this work, we introduce a practical incremental method to train deep CTR models, which consists of three decoupled modules (namely, data, feature and model module). Our method can achieve comparable performance to the conventional batch mode training with much better training efficiency. We conduct extensive experiments on a public benchmark and a private dataset to demonstrate the effectiveness of our proposed method.

cs.IR↗

arXiv2020

Personalized Re-ranking for Improving Diversity in Live Recommender Systems

Users of industrial recommender systems are normally suggesteda list of items at one time. Ideally, such list-wise recommendationshould provide diverse and relevant options to the users. However, in practice, list-wise recommendation is implemented as top-N recommendation. Top-N recommendation selects the first N items from candidates to display. The list is generated by a ranking function, which is learned from labeled data to optimize accuracy.However, top-N recommendation may lead to suboptimal, as it focuses on accuracy of each individual item independently and overlooks mutual influence between items. Therefore, we propose a personalized re-ranking model for improving diversity of the recommendation list in real recommender systems. The proposed re-ranking model can be easily deployed as a follow-up component after any existing ranking function. The re-ranking model improves the diversity by employing personalized Determinental Point Process (DPP). DPP has been applied in some recommender systems to improve the diversity and increase the user engagement.However, DPP does not take into account the fact that users may have individual propensities to the diversity. To overcome such limitation, our re-ranking model proposes a personalized DPP to model the trade-off between accuracy and diversity for each individual user. We implement and deploy the personalized DPP model on alarge scale industrial recommender system. Experimental results on both offline and online demonstrate the efficiency of our proposed re-ranking model.

cs.IR↗

arXiv2020

Uncovering Download Fraud Activities in Mobile App Markets

Download fraud is a prevalent threat in mobile App markets, where fraudsters manipulate the number of downloads of Apps via various cheating approaches. Purchased fake downloads can mislead recommendation and search algorithms and further lead to bad user experience in App markets. In this paper, we investigate download fraud problem based on a company's App Market, which is one of the most popular Android App markets. We release a honeypot App on the App Market and purchase fake downloads from fraudster agents to track fraud activities in the wild. Based on our interaction with the fraudsters, we categorize download fraud activities into three types according to their intentions: boosting front end downloads, optimizing App search ranking, and enhancing user acquisition&retention rate. For the download fraud aimed at optimizing App search ranking, we select, evaluate, and validate several features in identifying fake downloads based on billions of download data. To get a comprehensive understanding of download fraud, we further gather stances of App marketers, fraudster agencies, and market operators on download fraud. The followed analysis and suggestions shed light on the ways to mitigate download fraud in App markets and other social platforms. To the best of our knowledge, this is the first work that investigates the download fraud problem in mobile App markets.

cs.CR↗

arXiv2019

Two-dimensional ferromagnetic-ferroelectric multiferroics in violation of the d0 rule

Contribution of d-electron to ferroelectricity of type-II multiferroics causes strong magneto-electric coupling and distinguishes them from the conventional type-I multiferroics. However, their therein polarization is too small because the ferroelectricity is merely a derivative from the magnetic order. Here we report a new class of multiferroic materials, monolayer VOX2 (X = Cl, Br, and I), which combine the advantages of type-I and type-II multiferroics. Both ferroelectricity and magnetism arise from the same V cation, where the filled d-orbital is perpendicular to an a priori ferroelectric polarization and thus poses no hindrance to ferroelectricity, indicating a violation of the usual d0 rule. This makes the combination of large polarizations and strong magneto-electric coupling possible. Our findings not only add new ferromagnetic-ferroelectric multiferroics, but also point to a unique mechanism to engineer multiferroics.

cond-mat.mtrl-sci↗

arXiv2018

Raman Spectra and Strain Effects in Bismuth Oxychalcogenides

A new type of two-dimensional layered semiconductor with weak electrostatic but not van der Waals interlayer interactions, Bi2O2Se, has been recently synthesized, which shown excellent air stability and ultrahigh carrier mobility. Herein, we combined theoretical and experimental approaches to study the Raman spectra of Bi2O2Se and related bismuth oxychalcogenides (Bi2O2Te and Bi2O2S). The experimental peaks were fully consistent with the calculated results, and were successfully assigned. Bi2O2S was predicted to have more Raman-active modes due to its lower symmetry. The shift of the predicted frequencies of Raman active modes was also found to get softened as the interlayer interaction decreases from bulk to monolayer Bi2O2Se and Bi2O2Te. To reveal the strain effects on the Raman shifts, a universal theoretical equation was established based on the symmetry of Bi2O2Se and Bi2O2Te. It was predicted that the doubly degenerate modes split under in-plane uniaxial/shear strains. Under a rotated uniaxial strain, the changes of Raman shifts are anisotropic for degenerate modes although Bi2O2Se and Bi2O2Te were usually regarded as isotropic systems similar to graphene. This implies a novel method to identify the crystallographic orientation from Raman spectra under strain. These results have important consequences for the incorporation of 2D Bismuth oxychalcogenides into nanoelectronic devices.

cond-mat.mes-hall↗

arXiv2018

Angle-adjustable density field formulation for modeling crystalline microstructures

A continuum density-field formulation with particle-scale resolution is constructed to simultaneously incorporate the orientation dependence of interparticle interactions and the rotational invariance of the system, a fundamental but challenging issue in modeling structure and dynamics of a broad range of material systems across variable scales. This generalized phase field crystal type approach is based upon the complete expansion of particle direct correlation functions and the concept of isotropic tensors. Through applications to the modeling of various two- and three-dimensional crystalline structures, our study demonstrates the capability of bond angle control in this continuum field theory and its effects on the emergence of ordered phases, and provides a systematic way of tunable angle analysis for crystalline microstructures.

cond-mat.mtrl-sci↗

arXiv2018

Elastic constants of stressed and unstressed materials in the phase field crystal model

A general procedure to investigate the elastic response and calculate the elastic constants of stressed and unstressed materials through continuum field modeling, particularly the phase field crystal (PFC) models, is presented. It is found that for a complete description of system response to elastic deformation, the variations of all the quantities of lattice wave vectors, their density amplitudes (including the corresponding anisotropic variation and degeneracy breaking), the average atomic density, and system volume should be incorporated. The quantitative and qualitative results of elastic constant calculations highly depend on the physical interpretation of the density field used in the model, and also importantly, on the intrinsic pressure that usually pre-exists in the model system. A formulation based on thermodynamics is constructed to account for the effects caused by constant pre-existing stress during the homogeneous elastic deformation, through the introducing of a generalized Gibbs free energy and an effective finite strain tensor used for determining the elastic constants. The elastic properties of both solid and liquid states can be well produced by this unified approach, as demonstrated by an analysis for the liquid state and numerical evaluations for the bcc solid phase. The numerical calculations of bcc elastic constants and Poisson's ratio through this method generate results that are consistent with experimental conditions, and better match the data of bcc Fe given by molecular dynamics simulations as compared to previous work. The general theory developed here is applicable to the study of different types of stressed or unstressed material systems under elastic deformation.

cond-mat.mtrl-sci↗

arXiv2017

In-plane uniaxial strain in black phosphorus enables the identification of crystalline orientation

The identification of the crystalline axis of anisotropic black phosphorus (BP) is key for the study of its physical properties and for its optical and electronic applications. Herein, we show that by applying an in-plane uniaxial strain and measuring the changes of Raman shifts, the crystalline axis of BP can be reliably determined. The strain effects on Raman shifts are angle-dependent, which can be expressed as a combination of Raman responses under zigzag and armchair strains. Different from previous polarized optical spectroscopic means, the proposed method does not rely on the laser polarization, the excitation wavelength, the sample thickness and the material of the stretchable substrate. Besides, the effective strain applied on BP from the stretched substrate is estimated, showing that only 20% to 40% of the strain can be effectively transferred to BP flakes from polyethylene terephthalate substrates. Our method provides not only an effective and universal approach to identify the crystalline orientation of layered BP, but also a model to extract additional information in strain-related studies, and can be extended to other 2D anisotropic materials.

cond-mat.mtrl-sci↗