arXiv subjects

Bing Hu

Publications and source records attributed to Bing Hu.

At least 37 records · Page 2Linked to original sources

arXiv2025

Domain Knowledge Infused Conditional Generative Models for Accelerating Drug Discovery

The role of Artificial Intelligence (AI) is growing in every stage of drug development. Nevertheless, a major challenge in drug discovery AI remains: Drug pharmacokinetic (PK) and Drug-Target Interaction (DTI) datasets collected in different studies often exhibit limited overlap, creating data overlap sparsity. Thus, data curation becomes difficult, negatively impacting downstream research investigations in high-throughput screening, polypharmacy, and drug combination. We propose xImagand-DKI, a novel SMILES/Protein-to-Pharmacokinetic/DTI (SP2PKDTI) diffusion model capable of generating an array of PK and DTI target properties conditioned on SMILES and protein inputs that exhibit data overlap sparsity. We infuse additional molecular and genomic domain knowledge from the Gene Ontology (GO) and molecular fingerprints to further improve our model performance. We show that xImagand-DKI-generated synthetic PK data closely resemble real data univariate and bivariate distributions, and can adequately fill in gaps among PK and DTI datasets. As such, xImagand-DKI is a promising solution for data overlap sparsity and may improve performance for downstream drug discovery research tasks. Code available at: https://github.com/GenerativeDrugDiscovery/xImagand-DKI

q-bio.QM↗

arXiv2025

FlowMoE: A Scalable Pipeline Scheduling Framework for Distributed Mixture-of-Experts Training

The parameter size of modern large language models (LLMs) can be scaled up via the sparsely-activated Mixture-of-Experts (MoE) technique to avoid excessive increase of the computational costs. To further improve training efficiency, pipelining computation and communication has become a promising solution for distributed MoE training. However, existing work primarily focuses on scheduling tasks within the MoE layer, such as expert computing and all-to-all (A2A) communication, while neglecting other key operations including multi-head attention (MHA) computing, gating, and all-reduce communication. In this paper, we propose FlowMoE, a scalable framework for scheduling multi-type task pipelines. First, FlowMoE constructs a unified pipeline to consistently scheduling MHA computing, gating, expert computing, and A2A communication. Second, FlowMoE introduces a tensor chunk-based priority scheduling mechanism to overlap the all-reduce communication with all computing tasks. We implement FlowMoE as an adaptive and generic framework atop PyTorch. Extensive experiments with 675 typical MoE layers and four real-world MoE models across two GPU clusters demonstrate that our proposed FlowMoE framework outperforms state-of-the-art MoE training frameworks, reducing training time by 13%-57%, energy consumption by 10%-39%, and memory usage by 7%-32%.

cs.DC↗

arXiv2025

Drug Discovery SMILES-to-Pharmacokinetics Diffusion Models with Deep Molecular Understanding

Artificial intelligence (AI) is increasingly used in every stage of drug development. One challenge facing drug discovery AI is that drug pharmacokinetic (PK) datasets are often collected independently from each other, often with limited overlap, creating data overlap sparsity. Data sparsity makes data curation difficult for researchers looking to answer research questions in poly-pharmacy, drug combination research, and high-throughput screening. We propose Imagand, a novel SMILES-to-Pharmacokinetic (S2PK) diffusion model capable of generating an array of PK target properties conditioned on SMILES inputs. We show that Imagand-generated synthetic PK data closely resembles real data univariate and bivariate distributions, and improves performance for downstream tasks. Imagand is a promising solution for data overlap sparsity and allows researchers to efficiently generate ligand PK data for drug discovery research. Code is available at https://github.com/bing1100/Imagand.

q-bio.QM↗

arXiv2025

LION-FS: Fast & Slow Video-Language Thinker as Online Video Assistant

First-person video assistants are highly anticipated to enhance our daily lives through online video dialogue. However, existing online video assistants often sacrifice assistant efficacy for real-time efficiency by processing low-frame-rate videos with coarse-grained visual features.To overcome the trade-off between efficacy and efficiency, we propose "Fast & Slow Video-Language Thinker" as an onLIne videO assistaNt, LION-FS, achieving real-time, proactive, temporally accurate, and contextually precise responses. LION-FS adopts a two-stage optimization strategy: 1)Fast Path: Routing-Based Response Determination evaluates frame-by-frame whether an immediate response is necessary. To enhance response determination accuracy and handle higher frame-rate inputs efficiently, we employ Token Aggregation Routing to dynamically fuse spatiotemporal features without increasing token numbers, while utilizing Token Dropping Routing to eliminate redundant features. 2)Slow Path: Multi-granularity Keyframe Augmentation optimizes keyframes during response generation. To provide comprehensive and detailed responses beyond atomic actions constrained by training data, fine-grained spatial features and human-environment interaction features are extracted through multi-granular pooling. These features are further integrated into a meticulously designed multimodal Thinking Template to guide more precise response generation. Comprehensive evaluations on online video tasks demonstrate that LION-FS achieves state-of-the-art efficacy and efficiency.

cs.CV↗

arXiv2024

Bug In the Code Stack: Can LLMs Find Bugs in Large Python Code Stacks

Recent research in Needle-in-a-Haystack (NIAH) benchmarks has explored the capabilities of Large Language Models (LLMs) in retrieving contextual information from large text documents. However, as LLMs become increasingly integrated into software development processes, it is crucial to evaluate their performance in code-based environments. As LLMs are further developed for program synthesis, we need to ensure that LLMs can understand syntax and write syntactically correct code. As a step in ensuring LLMs understand syntax, LLMs can be evaluated in their ability to find and detect syntax bugs. Our benchmark, Bug In The Code Stack (BICS), is designed to assess the ability of LLMs to identify simple syntax bugs within large source code. Our findings reveal three key insights: (1) code-based environments pose significantly more challenge compared to text-based environments for retrieval tasks, (2) there is a substantial performance disparity among different models, and (3) there is a notable correlation between longer context lengths and performance degradation, though the extent of this degradation varies between models.

cs.AI↗

arXiv2024

The Solution for CVPR2024 Foundational Few-Shot Object Detection Challenge

This report introduces an enhanced method for the Foundational Few-Shot Object Detection (FSOD) task, leveraging the vision-language model (VLM) for object detection. However, on specific datasets, VLM may encounter the problem where the detected targets are misaligned with the target concepts of interest. This misalignment hinders the zero-shot performance of VLM and the application of fine-tuning methods based on pseudo-labels. To address this issue, we propose the VLM+ framework, which integrates the multimodal large language model (MM-LLM). Specifically, we use MM-LLM to generate a series of referential expressions for each category. Based on the VLM predictions and the given annotations, we select the best referential expression for each category by matching the maximum IoU. Subsequently, we use these referential expressions to generate pseudo-labels for all images in the training set and then combine them with the original labeled data to fine-tune the VLM. Additionally, we employ iterative pseudo-label generation and optimization to further enhance the performance of the VLM. Our approach achieve 32.56 mAP in the final test.

cs.CV↗

arXiv2024

Synthetic Data from Diffusion Models Improve Drug Discovery Prediction

Artificial intelligence (AI) is increasingly used in every stage of drug development. Continuing breakthroughs in AI-based methods for drug discovery require the creation, improvement, and refinement of drug discovery data. We posit a new data challenge that slows the advancement of drug discovery AI: datasets are often collected independently from each other, often with little overlap, creating data sparsity. Data sparsity makes data curation difficult for researchers looking to answer key research questions requiring values posed across multiple datasets. We propose a novel diffusion GNN model Syngand capable of generating ligand and pharmacokinetic data end-to-end. We show and provide a methodology for sampling pharmacokinetic data for existing ligands using our Syngand model. We show the initial promising results on the efficacy of the Syngand-generated synthetic target property data on downstream regression tasks with AqSolDB, LD50, and hERG central. Using our proposed model and methodology, researchers can easily generate synthetic ligand data to help them explore research questions that require data spanning multiple datasets.

cs.LG↗

arXiv2022

Mechanical control of physical properties in the van der Waals ferromagnet Cr2Ge2Te6 via application of electric current

Cr2Ge2Te6 is a van der Waals ferromagnet with a Curie temperature at 66 K. Here we report a swift change in the magnetic ground state upon application of small DC electric current, a giant yet anisotropic magnetoelectric effect, and a sharp, lattice-driven quantum switching manifested in the I-V characteristic of the bulk single-crystal Cr2Ge2Te6. At the heart of these observed phenomena is a newly uncovered, strongly anisotropic magnetoelastic coupling that enables strongly anisotropic responses of the lattice to application of electric current and/or magnetic field, thus the exotic phenomena in Cr2Ge2Te6. Such a rare mechanical tunability in the magnetic semiconductors promises tantalizing prospects for unique functional materials and devices.

cond-mat.mtrl-sci↗

arXiv2021

Graph Attention Network-based Multi-agent Reinforcement Learning for Slicing Resource Management in Dense Cellular Network

Network slicing (NS) management devotes to providing various services to meet distinct requirements over the same physical communication infrastructure and allocating resources on demands. Considering a dense cellular network scenario that contains several NS over multiple base stations (BSs), it remains challenging to design a proper real-time inter-slice resource management strategy, so as to cope with frequent BS handover and satisfy the fluctuations of distinct service requirements. In this paper, we propose to formulate this challenge as a multi-agent reinforcement learning (MARL) problem in which each BS represents an agent. Then, we leverage graph attention network (GAT) to strengthen the temporal and spatial cooperation between agents. Furthermore, we incorporate GAT into deep reinforcement learning (DRL) and correspondingly design an intelligent real-time inter-slice resource management strategy. More specially, we testify the universal effectiveness of GAT for advancing DRL in the multi-agent system, by applying GAT on the top of both the value-based method deep Q-network (DQN) and a combination of policy-based and value-based method advantage actor-critic (A2C). Finally, we verify the superiority of the GAT-based MARL algorithms through extensive simulations.

cs.MA↗

arXiv2021

The ground state in a proximity to a possible Kitaev spin liquid: An undistorted honeycomb iridate NaxIrO3 (0.60 < x < 0.80)

We report results of our study of a newly synthesized honeycomb iridate NaxIrO3 (0.60 < x < 0.80). Single-crystal NaxIrO3 adopts a honeycomb lattice noticeably without distortions and stacking disorder inherently existent in its sister compound Na2IrO3. The oxidation state of the Ir ion is a mixed valence state resulting from a majority Ir5+(5d4) ion and a minority Ir6+(5d3) ion. NaxIrO3 is a Mott insulator likely with a predominant pseudospin = 1 state. It exhibits an effective moment of 1.1 Bohr Magneton/Ir and a Curie-Weiss temperature of -19 K but with no discernable long-range order above 1 K. The physical behavior below 1 K features two prominent anomalies at Th = 0.9 K and Tl = 0.12 K in both the heat capacity and AC magnetic susceptibility. Intermediate between Th and Tl lies a pronounced temperature linearity of the heat capacity with a large slope of 77 mJ/mole K2, a feature expected for highly correlated metals but not at all for insulators. These results along with comparison drawn with the honeycomb lattices Na2IrO3 and (Na0.2Li0.8)2IrO3 point to an exotic ground state in a proximity to a possible Kitaev spin liquid.

cond-mat.str-el↗

arXiv2021

Strange electrical transport: Colossal magnetoresistance via avoiding fully polarized magnetization in ferrimagnetic insulator Mn3Si2Te6

Colossal magnetoresistance is of great fundamental and technological significance and exists mostly in the manganites and a few other materials. Here we report colossal magnetoresistance that is starkly different from that in all other materials. The stoichiometric Mn3Si2Te6 is an insulator featuring a ferrimagnetic transition at 78 K. The resistivity drops by 7 orders of magnitude with an applied magnetic field above 9 Tesla, leading to an insulator-metal transition at up to 130 K. However, the colossal magnetoresistance occurs only when the magnetic field is applied along the magnetic hard axis and is surprisingly absent when the magnetic field is applied along the magnetic easy axis where magnetization is fully saturated. The anisotropy field separating the easy and hard axes is 13 Tesla, unexpected for the Mn ions with nominally negligible orbital momentum and spin-orbit interactions. Double exchange and Jahn-Teller distortions that drive the hole-doped manganites do not exist in Mn3Si2Te6. The phenomena fit no existing models, suggesting a unique, intriguing type of electrical transport.

cond-mat.str-el↗

arXiv2020

Fortuitous partners of antiferromagnetic and Mott states in spin-orbit-coupled Sr2IrO4: A study of Sr2Ir1-xMxO4 (M=Fe or Co)

Sr2IrO4 is an archetypal spin-orbit-coupled Mott insulator with an antiferromagnetic state below 240 K. Here we report results of our study on single crystals of Sr2Ir1-xFexO4 (0<x<0.32) and Sr2Ir1-xCoxO4 (0<x<0.22). Fe doping retains the antiferromagnetic state but simultaneously precipitates an emergent metallic state whereas Co doping causes a rapid collapse of both the antiferromagnetic and Mott states, giving rise to a confined metallic state featuring a pronounced linearity of the basal-plane resistivity up to 700 K. The results indicate tetravalent Fe4+(3d4) ions in the intermediate spin state with S=1 and Co4+(3d5) ions in the high spin state with S=5/2 substituting for Ir4+(5d5) ions in Sr2IrO4, respectively. The effective magnetic moment closely tracks the Néel temperature as doping increases, suggesting that the spin state of the dopant predominately determines the magnetic properties in doped Sr2IrO4. Furthermore, all relevant properties including charge-carrier density (e.g., 1028/m3), Sommerfeld coefficient (e.g., 19 mJ/mole K2) and Wilson ratio (e.g., 2.6), consistently demonstrates a metallic state that is both robust and highly correlated in the two systems, arising from the percolation of bound states and the weakening of structural distortions. This study strongly suggests that the antiferromagnetic and Mott states merely coexist in a fortuitous manner in Sr2IrO4.

cond-mat.str-el↗

arXiv2020

Quest for New Quantum States via Field-Editing Technology

We report new quantum states in spin-orbit-coupled single crystals that are synthesized using a game-changing technology that "field-edits" crystal structures (borrowing from the phrase "genome editing") via application of magnetic field during crystal growth. This study is intended to fundamentally address a major challenge facing the research community today: A great deal of theoretical work predicting exotic states for strongly spin-orbit-coupled, correlated materials has thus far met very limited experimental confirmation. These conspicuous discrepancies are due chiefly to the extreme sensitivity of these materials to structural distortions. The results presented here demonstrate that the "field-edited" materials not only are much less distorted but also exhibit novel phenomena absent in their "non-edited" counterparts. The field-edited materials include an array of 4d and 5d transition metal oxides, and three representative materials presented here are Ba4Ir3O10, Ca2RuO4, and Sr2IrO4. This study provides an entirely new paradigm for discovery of new quantum states and materials otherwise unavailable.

cond-mat.str-el↗

arXiv2019

Nonequilibrium Orbital Transitions via Applied Electrical Current in Calcium Ruthenate

Simultaneous control of structural and physical properties via applied electrical current poses a key, new research topic and technological significance. Studying the spin-orbit-coupled antiferromagnet Ca2RuO4, with 3% Mn doping to weaken the violent first-order transition at 357 K for more robust measurements, we find that a small applied electrical current couples to the lattice by significantly reducing its orthorhombicity and octahedral rotations, concurrently diminishing the 125 K- antiferromagnetic transition and inducing a new, orbital order below 80 K. Our effort to establish a phase diagram reveals a critical regime near a current density of 0.15 A/cm2 that separates the vanishing antiferromagnetic order and the new orbital order. Further increasing current density (> 1 A/cm2) enhances competitions between relevant interactions in a metastable manner, leading to a peculiar glassy behavior above 80 K. The coupling between the lattice and nonequilibrium driven current is interpreted theoretically in terms of t2g orbital occupancies. The current-controlled lattice is the driving force of the observed novel phenomena.

cond-mat.str-el↗

arXiv2019

Restricted Deformable Convolution based Road Scene Semantic Segmentation Using Surround View Cameras

Understanding the surrounding environment of the vehicle is still one of the challenges for autonomous driving. This paper addresses 360-degree road scene semantic segmentation using surround view cameras, which are widely equipped in existing production cars. First, in order to address large distortion problem in the fisheye images, Restricted Deformable Convolution (RDC) is proposed for semantic segmentation, which can effectively model geometric transformations by learning the shapes of convolutional filters conditioned on the input feature map. Second, in order to obtain a large-scale training set of surround view images, a novel method called zoom augmentation is proposed to transform conventional images to fisheye images. Finally, an RDC based semantic segmentation model is built; the model is trained for real-world surround view images through a multi-task learning architecture by combining real-world images with transformed images. Experiments demonstrate the effectiveness of the RDC to handle images with large distortions, and that the proposed approach shows a good performance using surround view cameras with the help of the transformed images.

cs.CV↗

arXiv2019

A Novel ground state in a new dimer iridate Ba13Ir6O30 with Ir6+(5d3) Ions

We have synthesized and studied a new iridate, Ba13Ir6O30, with unusual Ir oxidation states: 2/3 Ir6+(5d3) ions and 1/3 Ir5+(5d4) ions. Its crystal structure features dimers of face-sharing IrO6 octahedra, and IrO6 monomers, that are linked via long, zigzag Ir-O-Ba-O-Ir pathways. Nevertheless, Ba13Ir6O30 exhibits two transitions at TN1 = 4.7 K and TN2 = 1.6 K. This magnetic order is accompanied by a huge Sommerfeld coefficient 200 mJ/mole K below TN2, signaling a coexisting frustrated/disordered state persisting down to at least 0.05 K. This iridate hosts unusually large Jeff=3/2 degrees of freedom, which is enabled by strong spin-orbit interactions (SOI) in the monomers with Ir6+ ions and a joint effect of molecular orbitals and SOI in the dimers occupied by Ir5+ and Ir6+ ions. Features displayed by the magnetization and heat capacity suggest that the combination of covalency, SOI and large effective spins leads to highly frustrated ferrimagnetic ordering, possibly into a skyrmion crystal, a novelty of this new high-spin iridate.

cond-mat.str-el↗

arXiv2016

Differential Reprogramming Based on Constructive Interference for Wireless Sensor Network

To improve the performance of reprogramming in wireless sensor network, we present a novel reprogramming structure and constructive interference-based dissemination protocol (CIDP) to transmit the patch through out the network fast and reliability. CIDP disseminates the patch, which is divided into several packets, to the network exploiting constructive interference. We evaluate our implementation of CIDP using simulation under different number of nodes. Our results show that CIDP disseminates the patch less than 4 milliseconds. In general, the probability of a node receives the complete patch as high as 99.99%.

cs.NI↗

arXiv2007

Singularity displacement in random speckle patterns of diffusive and localized waves: universality lost and regained

All random wave fields possess a network of phase singularities. We show that while the phase statistics within speckle patterns is generic, the statistics of the motion of phase singularities differs substantially for diffusive and localized waves. This reflects the changing wave interaction with the underlying modes of multiply-scattering systems. Universality is regained when the motion of phase singularities is charted against the phase excursion which reflects the variation of phase change across the speckle pattern. The evolution of speckle patterns can therefore be used to monitor internal motion in complex systems and to characterize the nature of wave propagation.

cond-mat.other↗