arXiv subjects

Hao Zhou

Publications and source records attributed to Hao Zhou.

At least 289 records · Page 16Linked to original sources

arXiv2024

Skim then Focus: Integrating Contextual and Fine-grained Views for Repetitive Action Counting

The key to action counting is accurately locating each video's repetitive actions. Instead of estimating the probability of each frame belonging to an action directly, we propose a dual-branch network, i.e., SkimFocusNet, working in a two-step manner. The model draws inspiration from empirical observations indicating that humans typically engage in coarse skimming of entire sequences to grasp the general action pattern initially, followed by a finer, frame-by-frame focus to determine if it aligns with the target action. Specifically, SkimFocusNet incorporates a skim branch and a focus branch. The skim branch scans the global contextual information throughout the sequence to identify potential target action for guidance. Subsequently, the focus branch utilizes the guidance to diligently identify repetitive actions using a long-short adaptive guidance (LSAG) block. Additionally, we have observed that videos in existing datasets often feature only one type of repetitive action, which inadequately represents real-world scenarios. To more accurately describe real-life situations, we establish the Multi-RepCount dataset, which includes videos containing multiple repetitive motions. On Multi-RepCount, our SkimFoucsNet can perform specified action counting, that is, to enable counting a particular action type by referencing an exemplary video. This capability substantially exhibits the robustness of our method. Extensive experiments demonstrate that SkimFocusNet achieves state-of-the-art performances with significant improvements. We also conduct a thorough ablation study to evaluate the network components. The source code will be published upon acceptance.

cs.CV↗

arXiv2024

Semi-Supervised Spoken Language Glossification

Spoken language glossification (SLG) aims to translate the spoken language text into the sign language gloss, i.e., a written record of sign language. In this work, we present a framework named $S$emi-$S$upervised $S$poken $L$anguage $G$lossification ($S^3$LG) for SLG. To tackle the bottleneck of limited parallel data in SLG, our $S^3$LG incorporates large-scale monolingual spoken language text into SLG training. The proposed framework follows the self-training structure that iteratively annotates and learns from pseudo labels. Considering the lexical similarity and syntactic difference between sign language and spoken language, our $S^3$LG adopts both the rule-based heuristic and model-based approach for auto-annotation. During training, we randomly mix these complementary synthetic datasets and mark their differences with a special token. As the synthetic data may be less quality, the $S^3$LG further leverages consistency regularization to reduce the negative impact of noise in the synthetic data. Extensive experiments are conducted on public benchmarks to demonstrate the effectiveness of the $S^3$LG. Our code is available at \url{https://github.com/yaohj11/S3LG}.

cs.CL↗

arXiv2024

An optical-ultraviolet flare with absolute AB magnitude of -39.4 detected in GRB 220101A

Hyperluminous optical-ultraviolet fares have been detected in gamma-ray bursts and the luminosity record was held by naked-eye event GRB 080319B. Such fares are widely attributed to internal shock or external reverse shock radiation. Here, with a new method developed to derive reliable photometry from saturated sources of Swift/UVOT, we carry out time-resolved analysis of the initial white-band 150s exposure of GRB 220101A, a burst at a redshift of 4.618, and report a rapidly evolving optical-ultraviolet fare with a high absolute AB magnitude of -39.4$\pm$0.2. In contrast to GRB 080319B, the temporal behaviour of this new fare does not trace the gamma-ray activity. Instead of either internal shocks or reverse shock, this extremely energetic optical-ultraviolet fare is most likely to originate from the refreshed shocks induced by the late-ejected extremely energetic material catching up with the earlier-launched decelerating outflow. This finding reveals the diverse origins of the extremely energetic optical-ultraviolet fares and demonstrates the necessity of high-time-resolution observations at early times.

astro-ph.HE↗

arXiv2024

StrucTexTv3: An Efficient Vision-Language Model for Text-rich Image Perception, Comprehension, and Beyond

Text-rich images have significant and extensive value, deeply integrated into various aspects of human life. Notably, both visual cues and linguistic symbols in text-rich images play crucial roles in information transmission but are accompanied by diverse challenges. Therefore, the efficient and effective understanding of text-rich images is a crucial litmus test for the capability of Vision-Language Models. We have crafted an efficient vision-language model, StrucTexTv3, tailored to tackle various intelligent tasks for text-rich images. The significant design of StrucTexTv3 is presented in the following aspects: Firstly, we adopt a combination of an effective multi-scale reduced visual transformer and a multi-granularity token sampler (MG-Sampler) as a visual token generator, successfully solving the challenges of high-resolution input and complex representation learning for text-rich images. Secondly, we enhance the perception and comprehension abilities of StrucTexTv3 through instruction learning, seamlessly integrating various text-oriented tasks into a unified framework. Thirdly, we have curated a comprehensive collection of high-quality text-rich images, abbreviated as TIM-30M, encompassing diverse scenarios like incidental scenes, office documents, web pages, and screenshots, thereby improving the robustness of our model. Our method achieved SOTA results in text-rich image perception tasks, and significantly improved performance in comprehension tasks. Among multimodal models with LLM decoder of approximately 1.8B parameters, it stands out as a leader, which also makes the deployment of edge devices feasible. In summary, the StrucTexTv3 model, featuring efficient structural design, outstanding performance, and broad adaptability, offers robust support for diverse intelligent application tasks involving text-rich images, thus exhibiting immense potential for widespread application.

cs.CV↗

arXiv2024

On Prompt-Driven Safeguarding for Large Language Models

Prepending model inputs with safety prompts is a common practice for safeguarding large language models (LLMs) against queries with harmful intents. However, the underlying working mechanisms of safety prompts have not been unraveled yet, restricting the possibility of automatically optimizing them to improve LLM safety. In this work, we investigate how LLMs' behavior (i.e., complying with or refusing user queries) is affected by safety prompts from the perspective of model representation. We find that in the representation space, the input queries are typically moved by safety prompts in a "higher-refusal" direction, in which models become more prone to refusing to provide assistance, even when the queries are harmless. On the other hand, LLMs are naturally capable of distinguishing harmful and harmless queries without safety prompts. Inspired by these findings, we propose a method for safety prompt optimization, namely DRO (Directed Representation Optimization). Treating a safety prompt as continuous, trainable embeddings, DRO learns to move the queries' representations along or opposite the refusal direction, depending on their harmfulness. Experiments with eight LLMs on out-of-domain and jailbreak benchmarks demonstrate that DRO remarkably improves the safeguarding performance of human-crafted safety prompts, without compromising the models' general performance.

cs.LG↗

arXiv2024

MolCRAFT: Structure-Based Drug Design in Continuous Parameter Space

Generative models for structure-based drug design (SBDD) have shown promising results in recent years. Existing works mainly focus on how to generate molecules with higher binding affinity, ignoring the feasibility prerequisites for generated 3D poses and resulting in false positives. We conduct thorough studies on key factors of ill-conformational problems when applying autoregressive methods and diffusion to SBDD, including mode collapse and hybrid continuous-discrete space. In this paper, we introduce MolCRAFT, the first SBDD model that operates in the continuous parameter space, together with a novel noise reduced sampling strategy. Empirical results show that our model consistently achieves superior performance in binding affinity with more stable 3D structure, demonstrating our ability to accurately model interatomic interactions. To our best knowledge, MolCRAFT is the first to achieve reference-level Vina Scores (-6.59 kcal/mol) with comparable molecular size, outperforming other strong baselines by a wide margin (-0.84 kcal/mol). Code is available at https://github.com/AlgoMole/MolCRAFT.

q-bio.BM↗

arXiv2024

Gradually Vanishing Gap in Prototypical Network for Unsupervised Domain Adaptation

Unsupervised domain adaptation (UDA) is a critical problem for transfer learning, which aims to transfer the semantic information from labeled source domain to unlabeled target domain. Recent advancements in UDA models have demonstrated significant generalization capabilities on the target domain. However, the generalization boundary of UDA models remains unclear. When the domain discrepancy is too large, the model can not preserve the distribution structure, leading to distribution collapse during the alignment. To address this challenge, we propose an efficient UDA framework named Gradually Vanishing Gap in Prototypical Network (GVG-PN), which achieves transfer learning from both global and local perspectives. From the global alignment standpoint, our model generates a domain-biased intermediate domain that helps preserve the distribution structures. By entangling cross-domain features, our model progressively reduces the risk of distribution collapse. However, only relying on global alignment is insufficient to preserve the distribution structure. To further enhance the inner relationships of features, we introduce the local perspective. We utilize the graph convolutional network (GCN) as an intuitive method to explore the internal relationships between features, ensuring the preservation of manifold structures and generating domain-biased prototypes. Additionally, we consider the discriminability of the inner relationships between features. We propose a pro-contrastive loss to enhance the discriminability at the prototype level by separating hard negative pairs. By incorporating both GCN and the pro-contrastive loss, our model fully explores fine-grained semantic relationships. Experiments on several UDA benchmarks validated that the proposed GVG-PN can clearly outperform the SOTA models.

cs.CV↗

arXiv2024

AnyFit: Controllable Virtual Try-on for Any Combination of Attire Across Any Scenario

While image-based virtual try-on has made significant strides, emerging approaches still fall short of delivering high-fidelity and robust fitting images across various scenarios, as their models suffer from issues of ill-fitted garment styles and quality degrading during the training process, not to mention the lack of support for various combinations of attire. Therefore, we first propose a lightweight, scalable, operator known as Hydra Block for attire combinations. This is achieved through a parallel attention mechanism that facilitates the feature injection of multiple garments from conditionally encoded branches into the main network. Secondly, to significantly enhance the model's robustness and expressiveness in real-world scenarios, we evolve its potential across diverse settings by synthesizing the residuals of multiple models, as well as implementing a mask region boost strategy to overcome the instability caused by information leakage in existing models. Equipped with the above design, AnyFit surpasses all baselines on high-resolution benchmarks and real-world data by a large gap, excelling in producing well-fitting garments replete with photorealistic and rich details. Furthermore, AnyFit's impressive performance on high-fidelity virtual try-ons in any scenario from any image, paves a new path for future research within the fashion community.

cs.CV↗

arXiv2024

Know in AdVance: Linear-Complexity Forecasting of Ad Campaign Performance with Evolving User Interest

Real-time Bidding (RTB) advertisers wish to \textit{know in advance} the expected cost and yield of ad campaigns to avoid trial-and-error expenses. However, Campaign Performance Forecasting (CPF), a sequence modeling task involving tens of thousands of ad auctions, poses challenges of evolving user interest, auction representation, and long context, making coarse-grained and static-modeling methods sub-optimal. We propose \textit{AdVance}, a time-aware framework that integrates local auction-level and global campaign-level modeling. User preference and fatigue are disentangled using a time-positioned sequence of clicked items and a concise vector of all displayed items. Cross-attention, conditioned on the fatigue vector, captures the dynamics of user interest toward each candidate ad. Bidders compete with each other, presenting a complete graph similar to the self-attention mechanism. Hence, we employ a Transformer Encoder to compress each auction into embedding by solving auxiliary tasks. These sequential embeddings are then summarized by a conditional state space model (SSM) to comprehend long-range dependencies while maintaining global linear complexity. Considering the irregular time intervals between auctions, we make SSM's parameters dependent on the current auction embedding and the time interval. We further condition SSM's global predictions on the accumulation of local results. Extensive evaluations and ablation studies demonstrate its superiority over state-of-the-art methods. AdVance has been deployed on the Tencent Advertising platform, and A/B tests show a remarkable 4.5\% uplift in Average Revenue per User (ARPU).

cs.IR↗

arXiv2024

GRB 231115A: a nearby Magnetar Giant Flare or a cosmic Short Gamma-Ray Burst?

There are two classes of gamma-ray transients with a duration shorter than 2 seconds. One consists of cosmic short Gamma-Ray Bursts (GRBs) taking place in the deep universe via the neutron star mergers, and the other is the magnetar giant flares (GFs) with energies of $\sim 10^{44}-10^{46}$ erg from ``nearby" galaxies. Though the magnetar GFs and the short GRBs have rather similar temporal and spectral properties, their energies ($E_{\rm γ,iso}$) are different by quite a few orders of magnitude and hence can be distinguished supposing the host galaxies have been robustly identified. The newly observed GRB 231115A has been widely discussed as a new GF event for its high probability of being associated with M82. Here we conduct a detailed analysis of its prompt emission observed by Fermi-GBM, and compare the parameters with existing observations. The prompt gamma-ray radiation properties of GRB 231115A, if associated with M82, nicely follow the $E_{\rm p,z}-E_{γ,\rm iso}$ relation of the GFs, where $E_{\rm p,z}$ is the peak energy of the gamma-ray spectrum after the redshift ($z$) correction. To be a short GRB, the reshift needs to be $\sim 1$. Though such a chance is low, the available X-ray/GeV observation upper limits are not stringent enough to further rule out this possibility. We have also discussed the prospect of convincingly establishing the magnetar origin of GRB 231115A-like events in the future.

astro-ph.HE↗

arXiv2024

Collage: Light-Weight Low-Precision Strategy for LLM Training

Large models training is plagued by the intense compute cost and limited hardware memory. A practical solution is low-precision representation but is troubled by loss in numerical accuracy and unstable training rendering the model less useful. We argue that low-precision floating points can perform well provided the error is properly compensated at the critical locations in the training process. We propose Collage which utilizes multi-component float representation in low-precision to accurately perform operations with numerical errors accounted. To understand the impact of imprecision to training, we propose a simple and novel metric which tracks the lost information during training as well as differentiates various precision strategies. Our method works with commonly used low-precision such as half-precision ($16$-bit floating points) and can be naturally extended to work with even lower precision such as $8$-bit. Experimental results show that pre-training using Collage removes the requirement of using $32$-bit floating-point copies of the model and attains similar/better training performance compared to $(16, 32)$-bit mixed-precision strategy, with up to $3.7\times$ speedup and $\sim 15\%$ to $23\%$ less memory usage in practice.

cs.LG↗

arXiv2024

Understanding the flat thermal conductivity of La2Zr2O7 at ultrahigh temperatures

Many crystals, such as lanthanum zirconate (La2Zr2O7), exhibit a flat temperature dependence of thermal conductivity at elevated temperatures. This phenomenon has recently been attributed to the inter-band phonon tunneling (or diffuson) contribution using different formalisms. However, the contributions of finite-temperature corrections (e.g., higher-order phonon-scattering, phonon renormalization, and phonon scattering cross-section softening effects) remain unclear. In this work, we predict and compare the thermal conductivity of La2Zr2O7 using three distinct first-principles methods. The first method is Green-Kubo molecular dynamics (MD) based on temperature-dependent machine learning interatomic potentials (MLIPs) trained from ab initio MD simulations, which successfully predict the flat trend at ultra-high temperatures. The second method is the Peierls Boltzmann transport equation (BTE), within the phonon particle framework, using phonon lifetime that includes all the finite-temperature corrections. Four-phonon scattering is found large but is cancelled by the phonon scattering cross-section softening effect. As a result, BTE with temperature corrections does not reproduce the flat thermal conductivity. The third method is Wigner formalism, which includes both phonon particle and wave contributions, which successfully reproduce the flat thermal conductivity. Diffuson and phonon contribute about 67% and 27% of thermal conductivity at 1800 K, respectively. The radiation contribution to thermal conductivity is around 6%. The scaling laws of the phonon, diffuson, radiation, and total thermal conductivity are found to be ~T-0.97, ~T0.43, ~T2.01, and ~T-0.40, respectively. This work clarifies the thermal transport mechanisms in La2Zr2O7 at ultra-high temperatures from different aspects.

cond-mat.mtrl-sci↗

arXiv2024

The Devil is in the Few Shots: Iterative Visual Knowledge Completion for Few-shot Learning

Contrastive Language-Image Pre-training (CLIP) has shown powerful zero-shot learning performance. Few-shot learning aims to further enhance the transfer capability of CLIP by giving few images in each class, aka 'few shots'. Most existing methods either implicitly learn from the few shots by incorporating learnable prompts or adapters, or explicitly embed them in a cache model for inference. However, the narrow distribution of few shots often contains incomplete class information, leading to biased visual knowledge with high risk of misclassification. To tackle this problem, recent methods propose to supplement visual knowledge by generative models or extra databases, which can be costly and time-consuming. In this paper, we propose an Iterative Visual Knowledge CompLetion (KCL) method to complement visual knowledge by properly taking advantages of unlabeled samples without access to any auxiliary or synthetic data. Specifically, KCL first measures the similarities between unlabeled samples and each category. Then, the samples with top confidence to each category is selected and collected by a designed confidence criterion. Finally, the collected samples are treated as labeled ones and added to few shots to jointly re-estimate the remaining unlabeled ones. The above procedures will be repeated for a certain number of iterations with more and more samples being collected until convergence, ensuring a progressive and robust knowledge completion process. Extensive experiments on 11 benchmark datasets demonstrate the effectiveness and efficiency of KCL as a plug-and-play module under both few-shot and zero-shot learning settings. Code is available at https://github.com/Mark-Sky/KCL.

cs.CV↗

arXiv2024

Eliciting the Translation Ability of Large Language Models via Multilingual Finetuning with Translation Instructions

Large-scale Pretrained Language Models (LLMs), such as ChatGPT and GPT4, have shown strong abilities in multilingual translations, without being explicitly trained on parallel corpora. It is interesting how the LLMs obtain their ability to carry out translation instructions for different languages. In this paper, we present a detailed analysis by finetuning a multilingual pretrained language model, XGLM-7B, to perform multilingual translation following given instructions. Firstly, we show that multilingual LLMs have stronger translation abilities than previously demonstrated. For a certain language, the performance depends on its similarity to English and the amount of data used in the pretraining phase. Secondly, we find that LLMs' ability to carry out translation instructions relies on the understanding of translation instructions and the alignment among different languages. With multilingual finetuning, LLMs could learn to perform the translation task well even for those language pairs unseen during the instruction tuning phase.

cs.CL↗

arXiv2024

Variance-reduced Zeroth-Order Methods for Fine-Tuning Language Models

Fine-tuning language models (LMs) has demonstrated success in a wide array of downstream tasks. However, as LMs are scaled up, the memory requirements for backpropagation become prohibitively high. Zeroth-order (ZO) optimization methods can leverage memory-efficient forward passes to estimate gradients. More recently, MeZO, an adaptation of ZO-SGD, has been shown to consistently outperform zero-shot and in-context learning when combined with suitable task prompts. In this work, we couple ZO methods with variance reduction techniques to enhance stability and convergence for inference-based LM fine-tuning. We introduce Memory-Efficient Zeroth-Order Stochastic Variance-Reduced Gradient (MeZO-SVRG) and demonstrate its efficacy across multiple LM fine-tuning tasks, eliminating the reliance on task-specific prompts. Evaluated across a range of both masked and autoregressive LMs on benchmark GLUE tasks, MeZO-SVRG outperforms MeZO with up to 20% increase in test accuracies in both full- and partial-parameter fine-tuning settings. MeZO-SVRG benefits from reduced computation time as it often surpasses MeZO's peak test accuracy with a $2\times$ reduction in GPU-hours. MeZO-SVRG significantly reduces the required memory footprint compared to first-order SGD, i.e. by $2\times$ for autoregressive models. Our experiments highlight that MeZO-SVRG's memory savings progressively improve compared to SGD with larger batch sizes.

cs.LG↗

arXiv2024

Towards Codable Watermarking for Injecting Multi-bits Information to LLMs

As large language models (LLMs) generate texts with increasing fluency and realism, there is a growing need to identify the source of texts to prevent the abuse of LLMs. Text watermarking techniques have proven reliable in distinguishing whether a text is generated by LLMs by injecting hidden patterns. However, we argue that existing LLM watermarking methods are encoding-inefficient and cannot flexibly meet the diverse information encoding needs (such as encoding model version, generation time, user id, etc.). In this work, we conduct the first systematic study on the topic of Codable Text Watermarking for LLMs (CTWL) that allows text watermarks to carry multi-bit customizable information. First of all, we study the taxonomy of LLM watermarking technologies and give a mathematical formulation for CTWL. Additionally, we provide a comprehensive evaluation system for CTWL: (1) watermarking success rate, (2) robustness against various corruptions, (3) coding rate of payload information, (4) encoding and decoding efficiency, (5) impacts on the quality of the generated text. To meet the requirements of these non-Pareto-improving metrics, we follow the most prominent vocabulary partition-based watermarking direction, and devise an advanced CTWL method named Balance-Marking. The core idea of our method is to use a proxy language model to split the vocabulary into probability-balanced parts, thereby effectively maintaining the quality of the watermarked text. Our code is available at https://github.com/lancopku/codable-watermarking-for-llm.

cs.CL↗

arXiv2024

HyperLLaVA: Dynamic Visual and Language Expert Tuning for Multimodal Large Language Models

Recent advancements indicate that scaling up Multimodal Large Language Models (MLLMs) effectively enhances performance on downstream multimodal tasks. The prevailing MLLM paradigm, \emph{e.g.}, LLaVA, transforms visual features into text-like tokens using a \emph{static} vision-language mapper, thereby enabling \emph{static} LLMs to develop the capability to comprehend visual information through visual instruction tuning. Although promising, the \emph{static} tuning strategy~\footnote{The static tuning refers to the trained model with static parameters.} that shares the same parameters may constrain performance across different downstream multimodal tasks. In light of this, we introduce HyperLLaVA, which involves adaptive tuning of the projector and LLM parameters, in conjunction with a dynamic visual expert and language expert, respectively. These experts are derived from HyperNetworks, which generates adaptive parameter shifts through visual and language guidance, enabling dynamic projector and LLM modeling in two-stage training. Our experiments demonstrate that our solution significantly surpasses LLaVA on existing MLLM benchmarks, including MME, MMBench, SEED-Bench, and LLaVA-Bench. ~\footnote{Our project is available on the link https://github.com/DCDmllm/HyperLLaVA}.

cs.AI↗

arXiv2024

DeepCRE: Transforming Drug R&D via AI-Driven Cross-drug Response Evaluation

The fields of therapeutic application and drug research and development (R&D) both face substantial challenges, i.e., the therapeutic domain calls for more treatment alternatives, while numerous promising pre-clinical drugs have failed in clinical trials. One of the reasons is the inadequacy of Cross-drug Response Evaluation (CRE) during the late stages of drug R&D. Although in-silico CRE models bring a promising solution, existing methodologies are restricted to early stages of drug R&D, such as target and cell-line levels, offering limited improvement to clinical success rates. Herein, we introduce DeepCRE, a pioneering AI model designed to predict CRE effectively in the late stages of drug R&D. DeepCRE outperforms the existing best models by achieving an average performance improvement of 17.7% in patient-level CRE, and a 5-fold increase in indication-level CRE, facilitating more accurate personalized treatment predictions and better pharmaceutical value assessment for indications, respectively. Furthermore, DeepCRE has identified a set of six drug candidates that show significantly greater effectiveness than a comparator set of two approved drugs in 5/8 colorectal cancer organoids. This demonstrates the capability of DeepCRE to systematically uncover a spectrum of drug candidates with enhanced therapeutic effects, highlighting its potential to transform drug R&D.

cs.AI↗