arXiv subjects

Jiahui Chen

Publications and source records attributed to Jiahui Chen.

At least 55 records · Page 3Linked to original sources

arXiv2025

UniSearch: Rethinking Search System with a Unified Generative Architecture

Modern search systems play a crucial role in facilitating information acquisition. Traditional search engines typically rely on a cascaded architecture, where results are retrieved through recall, pre-ranking, and ranking stages. The complexity of designing and maintaining multiple modules makes it difficult to achieve holistic performance gains. Recent advances in generative recommendation have motivated the exploration of unified generative search as an alternative. However, existing approaches are not genuinely end-to-end: they typically train an item encoder to tokenize candidates first and then optimize a generator separately, leading to objective inconsistency and limited generalization. To address these limitations, we propose UniSearch, a unified generative search framework for Kuaishou Search. UniSearch replaces the cascaded pipeline with an end-to-end architecture that integrates a Search Generator and a Video Encoder. The Generator produces semantic identifiers of relevant items given a user query, while the Video Encoder learns latent item embeddings and provides their tokenized representations. A unified training framework jointly optimizes both components, enabling mutual enhancement and improving representation quality and generation accuracy. Furthermore, we introduce Search Preference Optimization (SPO), which leverages a reward model and real user feedback to better align generation with user preferences. Extensive experiments on industrial-scale datasets, together with online A/B testing in both short-video and live search scenarios, demonstrate the strong effectiveness and deployment potential of UniSearch. Notably, its deployment in live search yields the largest single-experiment improvement in recent years of our product's history, highlighting its practical value for real-world applications.

cs.IR↗

arXiv2025

Beyond Interpretability: Exploring the Comprehensibility of Adaptive Video Streaming through Large Language Models

Over the past decade, adaptive video streaming technology has witnessed significant advancements, particularly driven by the rapid evolution of deep learning techniques. However, the black-box nature of deep learning algorithms presents challenges for developers in understanding decision-making processes and optimizing for specific application scenarios. Although existing research has enhanced algorithm interpretability through decision tree conversion, interpretability does not directly equate to developers' subjective comprehensibility. To address this challenge, we introduce \texttt{ComTree}, the first bitrate adaptation algorithm generation framework that considers comprehensibility. The framework initially generates the complete set of decision trees that meet performance requirements, then leverages large language models to evaluate these trees for developer comprehensibility, ultimately selecting solutions that best facilitate human understanding and enhancement. Experimental results demonstrate that \texttt{ComTree} significantly improves comprehensibility while maintaining competitive performance, showing potential for further advancement. The source code is available at https://github.com/thu-media/ComTree.

cs.MM↗

arXiv2025

3D Vision-tactile Reconstruction from Infrared and Visible Images for Robotic Fine-grained Tactile Perception

To achieve human-like haptic perception in anthropomorphic grippers, the compliant sensing surfaces of vision tactile sensor (VTS) must evolve from conventional planar configurations to biomimetically curved topographies with continuous surface gradients. However, planar VTSs have challenges when extended to curved surfaces, including insufficient lighting of surfaces, blurring in reconstruction, and complex spatial boundary conditions for surface structures. With an end goal of constructing a human-like fingertip, our research (i) develops GelSplitter3D by expanding imaging channels with a prism and a near-infrared (NIR) camera, (ii) proposes a photometric stereo neural network with a CAD-based normal ground truth generation method to calibrate tactile geometry, and (iii) devises a normal integration method with boundary constraints of depth prior information to correcting the cumulative error of surface integrals. We demonstrate better tactile sensing performance, a 40$\%$ improvement in normal estimation accuracy, and the benefits of sensor shapes in grasping and manipulation tasks.

cs.RO↗

arXiv2025

Simulation of the Background from $^{13}$C$(α, n)^{16}$O Reaction in the JUNO Scintillator

Large-scale organic liquid scintillator detectors are highly efficient in the detection of MeV-scale electron antineutrinos. These signal events can be detected through inverse beta decay on protons, which produce a positron accompanied by a neutron. A noteworthy background for antineutrinos coming from nuclear power reactors and from the depths of the Earth (geoneutrinos) is generated by ($α, n$) reactions. In organic liquid scintillator detectors, $α$ particles emitted from intrinsic contaminants such as $^{238}$U, $^{232}$Th, and $^{210}$Pb/$^{210}$Po, can be captured on $^{13}$C nuclei, followed by the emission of a MeV-scale neutron. Three distinct interaction mechanisms can produce prompt energy depositions preceding the delayed neutron capture, leading to a pair of events correlated in space and time within the detector. Thus, ($α, n$) reactions represent an indistinguishable background in liquid scintillator-based antineutrino detectors, where their expected rate and energy spectrum are typically evaluated via Monte Carlo simulations. This work presents results from the open-source SaG4n software, used to calculate the expected energy depositions from the neutron and any associated de-excitation products. Also simulated is a detailed detector response to these interactions, using a dedicated Geant4-based simulation software from the JUNO experiment. An expected measurable $^{13}$C$(α, n)^{16}$O event rate and reconstructed prompt energy spectrum with associated uncertainties, are presented in the context of JUNO, however, the methods and results are applicable and relevant to other organic liquid scintillator neutrino detectors.

physics.ins-det↗

arXiv2025

Dimensionality reduction for k-means clustering of large-scale influenza mutation datasets

Viral mutations pose significant threats to public health by increasing infectivity, strengthening vaccine resistance, and altering disease severity. To track these evolving patterns, agencies like the CDC annually evaluate thousands of virus strains, underscoring the urgent need to understand viral mutagenesis and evolution in depth. In this study, we integrate genomic analysis, clustering, and three leading dimensionality reduction approaches, namely, principal component analysis (PCA), t-distributed stochastic neighbor embedding (t-SNE), and uniform manifold approximation and projection (UMAP)-to investigate the effects of COVID-19 on influenza virus propagation. By applying these methods to extensive pre- and post-pandemic influenza datasets, we reveal how selective pressures during the pandemic have influenced the diversity of influenza genetics. Our findings indicate that combining robust dimension reduction with clustering yields critical insights into the complex dynamics of viral mutation, informing both future research directions and strategies for public health intervention.

q-bio.PE↗

arXiv2025

Audio-driven Gesture Generation via Deviation Feature in the Latent Space

Gestures are essential for enhancing co-speech communication, offering visual emphasis and complementing verbal interactions. While prior work has concentrated on point-level motion or fully supervised data-driven methods, we focus on co-speech gestures, advocating for weakly supervised learning and pixel-level motion deviations. We introduce a weakly supervised framework that learns latent representation deviations, tailored for co-speech gesture video generation. Our approach employs a diffusion model to integrate latent motion features, enabling more precise and nuanced gesture representation. By leveraging weakly supervised deviations in latent space, we effectively generate hand gestures and mouth movements, crucial for realistic video production. Experiments show our method significantly improves video quality, surpassing current state-of-the-art techniques.

cs.CV↗

arXiv2025

Augmented Conditioning Is Enough For Effective Training Image Generation

Image generation abilities of text-to-image diffusion models have significantly advanced, yielding highly photo-realistic images from descriptive text and increasing the viability of leveraging synthetic images to train computer vision models. To serve as effective training data, generated images must be highly realistic while also sufficiently diverse within the support of the target data distribution. Yet, state-of-the-art conditional image generation models have been primarily optimized for creative applications, prioritizing image realism and prompt adherence over conditional diversity. In this paper, we investigate how to improve the diversity of generated images with the goal of increasing their effectiveness to train downstream image classification models, without fine-tuning the image generation model. We find that conditioning the generation process on an augmented real image and text prompt produces generations that serve as effective synthetic datasets for downstream training. Conditioning on real training images contextualizes the generation process to produce images that are in-domain with the real image distribution, while data augmentations introduce visual diversity that improves the performance of the downstream classifier. We validate augmentation-conditioning on a total of five established long-tail and few-shot image classification benchmarks and show that leveraging augmentations to condition the generation process results in consistent improvements over the state-of-the-art on the long-tailed benchmark and remarkable gains in extreme few-shot regimes of the remaining four benchmarks. These results constitute an important step towards effectively leveraging synthetic data for downstream training.

cs.CV↗

arXiv2025

Drug Resistance Predictions Based on a Directed Flag Transformer

The continuous evolution of the SARS-CoV-2 virus poses a significant challenge to global public health. Of particular concern is the potential resistance to the widely prescribed drug PAXLOVID, of which the main ingredient nirmatrelvir inhibits the viral main protease (Mpro). Here, we developed CAPTURE (direCted flAg laPlacian Transformer for drUg Resistance prEdictions) to analyze the effects of Mpro mutations on nirmatrelvir-Mpro binding affinities and identify potential drug-resistant mutations. CAPTURE combines a comprehensive mutation analysis with a resistance prediction module based on DFFormer-seq, which is a novel ensemble model that leverages a new Directed Flag Transformer and sequence embeddings from the protein and small-molecule-large-language models. Our analysis of the evolution of Mpro mutations revealed a progressive increase in mutation frequencies for residues near the binding site between May and December 2022, suggesting that the widespread use of PAXLOVID created a selective pressure that accelerated the evolution of drug-resistant variants. Applied to mutations at the nirmatrelvir-Mpro binding site, CAPTURE identified several potential resistance mutations, including H172Y and F140L, which have been experimentally confirmed, as well as five other mutations that await experimental verification. CAPTURE evaluation in a limited experimental data set on Mpro mutants gives a recall of 57\% and a precision of 71\% for predicting potential drug-resistant mutations. Our work establishes a powerful new framework for predicting drug-resistant mutations and real-time viral surveillance. The insights also guide the rational design of more resilient next-generation therapeutics.

q-bio.QM↗

arXiv2024

PLD-Tree: Persistent Laplacian Decision Tree for Protein-Protein Binding Free Energy Prediction

Recent advances in topology-based modeling have accelerated progress in physical modeling and molecular studies, including applications to protein-ligand binding affinity. In this work, we introduce the Persistent Laplacian Decision Tree (PLD-Tree), a novel method designed to address the challenging task of predicting protein-protein interaction (PPI) affinities. PLD-Tree focuses on protein chains at binding interfaces and employs the persistent Laplacian to capture topological invariants reflecting critical inter-protein interactions. These topological descriptors, derived from persistent homology, are further enhanced by incorporating evolutionary scale modeling (ESM) from a large language model to integrate sequence-based information. We validate PLD-Tree on two benchmark datasets-PDBbind V2020 and SKEMPI v2 demonstrating a correlation coefficient ($R_p$) of 0.83 under the sophisticated leave-out-protein-out cross-validation. Notably, our approach outperforms all reported state-of-the-art methods on these datasets. These results underscore the power of integrating machine learning techniques with topology-based descriptors for molecular docking and virtual screening, providing a robust and accurate framework for predicting protein-protein binding affinities.

q-bio.BM↗

arXiv2024

Text-to-Sticker: Style Tailoring Latent Diffusion Models for Human Expression

We introduce Style Tailoring, a recipe to finetune Latent Diffusion Models (LDMs) in a distinct domain with high visual quality, prompt alignment and scene diversity. We choose sticker image generation as the target domain, as the images significantly differ from photorealistic samples typically generated by large-scale LDMs. We start with a competent text-to-image model, like Emu, and show that relying on prompt engineering with a photorealistic model to generate stickers leads to poor prompt alignment and scene diversity. To overcome these drawbacks, we first finetune Emu on millions of sticker-like images collected using weak supervision to elicit diversity. Next, we curate human-in-the-loop (HITL) Alignment and Style datasets from model generations, and finetune to improve prompt alignment and style alignment respectively. Sequential finetuning on these datasets poses a tradeoff between better style alignment and prompt alignment gains. To address this tradeoff, we propose a novel fine-tuning method called Style Tailoring, which jointly fits the content and style distribution and achieves best tradeoff. Evaluation results show our method improves visual quality by 14%, prompt alignment by 16.2% and scene diversity by 15.3%, compared to prompt engineering the base Emu model for stickers generation.

cs.CV↗

arXiv2024

Self-Supervised Learning of Deviation in Latent Representation for Co-speech Gesture Video Generation

Gestures are pivotal in enhancing co-speech communication. While recent works have mostly focused on point-level motion transformation or fully supervised motion representations through data-driven approaches, we explore the representation of gestures in co-speech, with a focus on self-supervised representation and pixel-level motion deviation, utilizing a diffusion model which incorporates latent motion features. Our approach leverages self-supervised deviation in latent representation to facilitate hand gestures generation, which are crucial for generating realistic gesture videos. Results of our first experiment demonstrate that our method enhances the quality of generated videos, with an improvement from 2.7 to 4.5% for FGD, DIV, and FVD, and 8.1% for PSNR, 2.5% for SSIM over the current state-of-the-art methods.

cs.CV↗

arXiv2024

LW-DETR: A Transformer Replacement to YOLO for Real-Time Detection

In this paper, we present a light-weight detection transformer, LW-DETR, which outperforms YOLOs for real-time object detection. The architecture is a simple stack of a ViT encoder, a projector, and a shallow DETR decoder. Our approach leverages recent advanced techniques, such as training-effective techniques, e.g., improved loss and pretraining, and interleaved window and global attentions for reducing the ViT encoder complexity. We improve the ViT encoder by aggregating multi-level feature maps, and the intermediate and final feature maps in the ViT encoder, forming richer feature maps, and introduce window-major feature map organization for improving the efficiency of interleaved attention computation. Experimental results demonstrate that the proposed approach is superior over existing real-time detectors, e.g., YOLO and its variants, on COCO and other benchmark datasets. Code and models are available at (https://github.com/Atten4Vis/LW-DETR).

cs.CV↗

arXiv2024

PepHarmony: A Multi-View Contrastive Learning Framework for Integrated Sequence and Structure-Based Peptide Encoding

Recent advances in protein language models have catalyzed significant progress in peptide sequence representation. Despite extensive exploration in this field, pre-trained models tailored for peptide-specific needs remain largely unaddressed due to the difficulty in capturing the complex and sometimes unstable structures of peptides. This study introduces a novel multi-view contrastive learning framework PepHarmony for the sequence-based peptide encoding task. PepHarmony innovatively combines both sequence- and structure-level information into a sequence-level encoding module through contrastive learning. We carefully select datasets from the Protein Data Bank (PDB) and AlphaFold database to encompass a broad spectrum of peptide sequences and structures. The experimental data highlights PepHarmony's exceptional capability in capturing the intricate relationship between peptide sequences and structures compared with the baseline and fine-tuned models. The robustness of our model is confirmed through extensive ablation studies, which emphasize the crucial roles of contrastive loss and strategic data sorting in enhancing predictive performance. The proposed PepHarmony framework serves as a notable contribution to peptide representations, and offers valuable insights for future applications in peptide drug discovery and peptide engineering. We have made all the source code utilized in this study publicly accessible via GitHub at https://github.com/zhangruochi/PepHarmony or http://www.healthinformaticslab.org/supp/.

cs.LG↗

arXiv2023

Poisson-Boltzmann based machine learning (PBML) model for electrostatic analysis

Electrostatics is of paramount importance to chemistry, physics, biology, and medicine. The Poisson-Boltzmann (PB) theory is a primary model for electrostatic analysis. However, it is highly challenging to compute accurate PB electrostatic solvation free energies for macromolecules due to the nonlinearity, dielectric jumps, charge singularity , and geometric complexity associated with the PB equation. The present work introduces a PB based machine learning (PBML) model for biomolecular electrostatic analysis. Trained with the second-order accurate MIBPB solver, the proposed PBML model is found to be more accurate and faster than several eminent PB solvers in electrostatic analysis. The proposed PBML model can provide highly accurate PB electrostatic solvation free energy of new biomolecules or new conformations generated by molecular dynamics with much reduced computational cost.

physics.chem-ph↗

arXiv2023

Integration of persistent Laplacian and pre-trained transformer for protein solubility changes upon mutation

Protein mutations can significantly influence protein solubility, which results in altered protein functions and leads to various diseases. Despite of tremendous effort, machine learning prediction of protein solubility changes upon mutation remains a challenging task as indicated by the poor scores of normalized Correct Prediction Ratio (CPR). Part of the challenge stems from the fact that there is no three-dimensional (3D) structures for the wild-type and mutant proteins. This work integrates persistent Laplacians and pre-trained Transformer for the task. The Transformer, pretrained with hunderds of millions of protein sequences, embeds wild-type and mutant sequences, while persistent Laplacians track the topological invariant change and homotopic shape evolution induced by mutations in 3D protein structures, which are rendered from AlphaFold2. The resulting machine learning model was trained on an extensive data set labeled with three solubility types. Our model outperforms all existing predictive methods and improves the state-of-the-art up to 15%.

q-bio.BM↗

arXiv2023

Combinatorial and Hodge Laplacians: Similarity and Difference

As key subjects in spectral geometry and combinatorial graph theory respectively, the (continuous) Hodge Laplacian and the combinatorial Laplacian share similarities in revealing the topological dimension and geometric shape of data and in their realization of diffusion and minimization of harmonic measures. It is believed that they also both associate with vector calculus, through the gradient, curl, and divergence, as argued in the popular usage of "Hodge Laplacians on graphs" in the literature. Nevertheless, these Laplacians are intrinsically different in their domains of definitions and applicability to specific data formats, hindering any in-depth comparison of the two approaches. To facilitate the comparison and bridge the gap between the combinatorial Laplacian and Hodge Laplacian for the discretization of continuous manifolds with boundary, we further introduce Boundary-Induced Graph (BIG) Laplacians using tools from Discrete Exterior Calculus (DEC). BIG Laplacians are defined on discrete domains with appropriate boundary conditions to characterize the topology and shape of data. The similarities and differences of the combinatorial Laplacian, BIG Laplacian, and Hodge Laplacian are then examined. Through an Eulerian representation of 3D domains as level-set functions on regular grids, we show experimentally the conditions for the convergence of BIG Laplacian eigenvalues to those of the Hodge Laplacian for elementary shapes.

math.DG↗

arXiv2023

FedDef: Defense Against Gradient Leakage in Federated Learning-based Network Intrusion Detection Systems

Deep learning (DL) methods have been widely applied to anomaly-based network intrusion detection system (NIDS) to detect malicious traffic. To expand the usage scenarios of DL-based methods, federated learning (FL) allows multiple users to train a global model on the basis of respecting individual data privacy. However, it has not yet been systematically evaluated how robust FL-based NIDSs are against existing privacy attacks under existing defenses. To address this issue, we propose two privacy evaluation metrics designed for FL-based NIDSs, including (1) privacy score that evaluates the similarity between the original and recovered traffic features using reconstruction attacks, and (2) evasion rate against NIDSs using adversarial attack with the recovered traffic. We conduct experiments to illustrate that existing defenses provide little protection and the corresponding adversarial traffic can even evade the SOTA NIDS Kitsune. To defend against such attacks and build a more robust FL-based NIDS, we further propose FedDef, a novel optimization-based input perturbation defense strategy with theoretical guarantee. It achieves both high utility by minimizing the gradient distance and strong privacy protection by maximizing the input distance. We experimentally evaluate four existing defenses on four datasets and show that our defense outperforms all the baselines in terms of privacy protection with up to 7 times higher privacy score, while maintaining model accuracy loss within 3% under optimal parameter combination.

cs.CR↗

arXiv2023

Towards a muon scattering tomography system for both low-Z and high-Z materials

Muon scattering tomography (MST) is a non-destructive technique to image various materials by utilizing cosmic ray muons as probes. A typical MST system with a two-fold track detectors is particularly effective in detecting high-$Z$ materials (e.g. nuclear materials), but difficult to recognize low-$Z$ materials (e.g. explosive materials). In this work, we present a concept of MST system to discriminate both low-$Z$ and high-$Z$ materials by extra measuring momentum of low-energy muons with a Cherenkov detector. A toy Monte Carlo simulation to describe detector responses and multiple scatterings of a muon tracking through materials is developed for statistical tests. Based on momentum-dependent track reconstruction and image reconstruction algorithm, we evaluate separation powers of different materials in the system. The results show that momentum measurement of low-energy muons and accurate track reconstruction can improve separation power of low-$Z$ materials significantly. This may enable the MST system to detect both low-$Z$ and high-$Z$ materials with cosmic ray muons in the whole energy range.

physics.ins-det↗