arXiv subjects

Chen Li

Publications and source records attributed to Chen Li.

At least 307 records · Page 17Linked to original sources

arXiv2024

Tackling the Singularities at the Endpoints of Time Intervals in Diffusion Models

Most diffusion models assume that the reverse process adheres to a Gaussian distribution. However, this approximation has not been rigorously validated, especially at singularities, where t=0 and t=1. Improperly dealing with such singularities leads to an average brightness issue in applications, and limits the generation of images with extreme brightness or darkness. We primarily focus on tackling singularities from both theoretical and practical perspectives. Initially, we establish the error bounds for the reverse process approximation, and showcase its Gaussian characteristics at singularity time steps. Based on this theoretical insight, we confirm the singularity at t=1 is conditionally removable while it at t=0 is an inherent property. Upon these significant conclusions, we propose a novel plug-and-play method SingDiffusion to address the initial singular time step sampling, which not only effectively resolves the average brightness issue for a wide range of diffusion models without extra training efforts, but also enhances their generation capability in achieving notable lower FID scores.

cs.CV↗

arXiv2024

Anomalous Neutron Nuclear-Magnetic Interference Spectroscopy

The electron-phonon interaction plays a critical role in materials electrical, thermal, optical, and superconducting properties. However, measuring the phonon mode-resolved electron-phonon interaction has been challenging. Here we propose neutron-scattering-based Anomalous Neutron nUclear-Magnetic Interference Spectroscopy (ANUBIS), where the co-existence of neutron nuclear scattering and magnetic scattering leads to anomalous dynamical structure factor under the presence of the electron-phonon interaction. Such anomalous structure factor is linear in electron-phonon coupling constant at the phonon wavevector, and is directly proportional to the momentum and energy-resolved dielectric function. The experimental configuration can be achieved using existing polarized inelastic neutron scattering setup, and an order-of-magnitude estimate shows the viability to observe the anomalous scattering signal is around $10^{-4}$ to $10^{-3}$ relative to phonon scattering, which is achievable at emerging neutron facilities. Our proposal offers an alternative neutron-based metrology to probe the crucial electronic properties.

cond-mat.mtrl-sci↗

arXiv2024

mPLUG-DocOwl 1.5: Unified Structure Learning for OCR-free Document Understanding

Structure information is critical for understanding the semantics of text-rich images, such as documents, tables, and charts. Existing Multimodal Large Language Models (MLLMs) for Visual Document Understanding are equipped with text recognition ability but lack general structure understanding abilities for text-rich document images. In this work, we emphasize the importance of structure information in Visual Document Understanding and propose the Unified Structure Learning to boost the performance of MLLMs. Our Unified Structure Learning comprises structure-aware parsing tasks and multi-grained text localization tasks across 5 domains: document, webpage, table, chart, and natural image. To better encode structure information, we design a simple and effective vision-to-text module H-Reducer, which can not only maintain the layout information but also reduce the length of visual features by merging horizontal adjacent patches through convolution, enabling the LLM to understand high-resolution images more efficiently. Furthermore, by constructing structure-aware text sequences and multi-grained pairs of texts and bounding boxes for publicly available text-rich images, we build a comprehensive training set DocStruct4M to support structure learning. Finally, we construct a small but high-quality reasoning tuning dataset DocReason25K to trigger the detailed explanation ability in the document domain. Our model DocOwl 1.5 achieves state-of-the-art performance on 10 visual document understanding benchmarks, improving the SOTA performance of MLLMs with a 7B LLM by more than 10 points in 5/10 benchmarks. Our codes, models, and datasets are publicly available at https://github.com/X-PLUG/mPLUG-DocOwl/tree/main/DocOwl1.5.

cs.CV↗

arXiv2024

Efficient Diffusion Training via Min-SNR Weighting Strategy

Denoising diffusion models have been a mainstream approach for image generation, however, training these models often suffers from slow convergence. In this paper, we discovered that the slow convergence is partly due to conflicting optimization directions between timesteps. To address this issue, we treat the diffusion training as a multi-task learning problem, and introduce a simple yet effective approach referred to as Min-SNR-$γ$. This method adapts loss weights of timesteps based on clamped signal-to-noise ratios, which effectively balances the conflicts among timesteps. Our results demonstrate a significant improvement in converging speed, 3.4$\times$ faster than previous weighting strategies. It is also more effective, achieving a new record FID score of 2.06 on the ImageNet $256\times256$ benchmark using smaller architectures than that employed in previous state-of-the-art. The code is available at https://github.com/TiankaiHang/Min-SNR-Diffusion-Training.

cs.CV↗

arXiv2024

Enhancing Multi-Hop Knowledge Graph Reasoning through Reward Shaping Techniques

In the realm of computational knowledge representation, Knowledge Graph Reasoning (KG-R) stands at the forefront of facilitating sophisticated inferential capabilities across multifarious domains. The quintessence of this research elucidates the employment of reinforcement learning (RL) strategies, notably the REINFORCE algorithm, to navigate the intricacies inherent in multi-hop KG-R. This investigation critically addresses the prevalent challenges introduced by the inherent incompleteness of Knowledge Graphs (KGs), which frequently results in erroneous inferential outcomes, manifesting as both false negatives and misleading positives. By partitioning the Unified Medical Language System (UMLS) benchmark dataset into rich and sparse subsets, we investigate the efficacy of pre-trained BERT embeddings and Prompt Learning methodologies to refine the reward shaping process. This approach not only enhances the precision of multi-hop KG-R but also sets a new precedent for future research in the field, aiming to improve the robustness and accuracy of knowledge inference within complex KG frameworks. Our work contributes a novel perspective to the discourse on KG reasoning, offering a methodological advancement that aligns with the academic rigor and scholarly aspirations of the Natural journal, promising to invigorate further advancements in the realm of computational knowledge representation.

cs.AI↗

arXiv2024

Stochastic Analysis of Touch-Tone Frequency Recognition in Two-Way Radio Systems for Dialed Telephone Number Identification

This paper focuses on recognizing dialed numbers in a touch-tone telephone system based on the Dual Tone MultiFrequency (DTMF) signaling technique with analysis of stochastic aspects during the noise and random duration of characters. Each dialed digit's acoustic profile is derived from a composite of two carrier frequencies, distinctly assigned to represent that digit. The identification of each digit is achieved by pinpointing the frequency pair with the highest energy or amplitude in its spectral output, utilizing the Discrete-Time Fourier Transform (DTFT). This analysis includes simulations that illustrate the effects of introducing stochastic variations during the "Mark" and "Space" intervals of the decoding process, offering insights into the technique's efficacy and the impact of random temporal fluctuations. This will reduce the accuracy of decoder decoding and lower the SNR of the system.

eess.SP↗

arXiv2024

Common 7B Language Models Already Possess Strong Math Capabilities

Mathematical capabilities were previously believed to emerge in common language models only at a very large scale or require extensive math-related pre-training. This paper shows that the LLaMA-2 7B model with common pre-training already exhibits strong mathematical abilities, as evidenced by its impressive accuracy of 97.7% and 72.0% on the GSM8K and MATH benchmarks, respectively, when selecting the best response from 256 random generations. The primary issue with the current base model is the difficulty in consistently eliciting its inherent mathematical capabilities. Notably, the accuracy for the first answer drops to 49.5% and 7.9% on the GSM8K and MATH benchmarks, respectively. We find that simply scaling up the SFT data can significantly enhance the reliability of generating correct answers. However, the potential for extensive scaling is constrained by the scarcity of publicly available math questions. To overcome this limitation, we employ synthetic data, which proves to be nearly as effective as real data and shows no clear saturation when scaled up to approximately one million samples. This straightforward approach achieves an accuracy of 82.6% on GSM8K and 40.6% on MATH using LLaMA-2 7B models, surpassing previous models by 14.2% and 20.8%, respectively. We also provide insights into scaling behaviors across different reasoning complexities and error types.

cs.CL↗

arXiv2024

Temporal Talbot interferometer of strongly interacting molecular Bose-Einstein condensate

Talbot interferometer, as a periodic reproduction of momentum distribution in the time domain, finds significant applications in multiple research. The inter-particle interactions during the diffraction and interference process introduce numerous many-body physics problems, leading to unconventional interference characteristics. This work investigates both experimentally and theoretically the influence of interaction in a Talbot interferometer with a $^{6}\rm Li_2$ molecular Bose-Einstein condensate in a one-dimensional optical lattice, with interaction strength directly tunable via magnetic Feshbach resonance. A clear dependence of the period and amplitude of signal revivals on the interaction strength can be observed. While interactions increase the decay rate of the signal and advance the revivals, we find that over a wide range of interactions, the Talbot interferometer remains highly effective over a certain evolutionary timescale, including the case of fractional Talbot interference. This work provides insight into the interplay between interaction and the coherence properties of a temporal Talbot interference in optical lattices, paving the way for research into quantum interference in strongly interacting systems.

cond-mat.quant-gas↗

arXiv2024

Coffee: Cost-Effective Edge Caching for 360 Degree Live Video Streaming

While live 360 degree video streaming delivers immersive viewing experience, it poses significant bandwidth and latency challenges for content delivery networks. Edge servers are expected to play an important role in facilitating live streaming of 360 degree videos. In this paper, we propose a novel predictive edge caching algorithm (Coffee) for live 360 degree video that employ collaborative FoV prediction and predictive tile prefetching to reduce bandwidth consumption, streaming cost and improve the streaming quality and robustness. Our light-weight caching algorithms exploit the unique tile consumption patterns of live 360 degree video streaming to achieve high tile caching gains. Through extensive experiments driven by real 360 degree video streaming traces, we demonstrate that edge caching algorithms specifically designed for live 360 degree video streaming can achieve high streaming cost savings with small edge cache space consumption. Coffee, guided by viewer FoV predictions, significantly reduces back-haul traffic up to 76% compared to state-of-the-art edge caching algorithms. Furthermore, we develop a transcoding-aware variant (TransCoffee) and evaluate it using comprehensive experiments, which demonstrate that TransCoffee can achieve 63\% lower cost compared to state-of-the-art transcoding-aware approaches.

cs.MM↗

arXiv2024

Bayesian Inference Accelerator for Spiking Neural Networks

Bayesian neural networks offer better estimates of model uncertainty compared to frequentist networks. However, inference involving Bayesian models requires multiple instantiations or sampling of the network parameters, requiring significant computational resources. Compared to traditional deep learning networks, spiking neural networks (SNNs) have the potential to reduce computational area and power, thanks to their event-driven and spike-based computational framework. Most works in literature either address frequentist SNN models or non-spiking Bayesian neural networks. In this work, we demonstrate an optimization framework for developing and implementing efficient Bayesian SNNs in hardware by additionally restricting network weights to be binary-valued to further decrease power and area consumption. We demonstrate accuracies comparable to Bayesian binary networks with full-precision Bernoulli parameters, while requiring up to $25\times$ less spikes than equivalent binary SNN implementations. We show the feasibility of the design by mapping it onto Zynq-7000, a lightweight SoC, and achieve a $6.5 \times$ improvement in GOPS/DSP while utilizing up to 30 times less power compared to the state-of-the-art.

cs.NE↗

arXiv2024

Multi-Granularity Information Interaction Framework for Incomplete Utterance Rewriting

Recent approaches in Incomplete Utterance Rewriting (IUR) fail to capture the source of important words, which is crucial to edit the incomplete utterance, and introduce words from irrelevant utterances. We propose a novel and effective multi-task information interaction framework including context selection, edit matrix construction, and relevance merging to capture the multi-granularity of semantic information. Benefiting from fetching the relevant utterance and figuring out the important words, our approach outperforms existing state-of-the-art models on two benchmark datasets Restoration-200K and CANAND in this field. Code will be provided on \url{https://github.com/yanmenxue/QR}.

cs.CL↗

arXiv2024

SwitchTab: Switched Autoencoders Are Effective Tabular Learners

Self-supervised representation learning methods have achieved significant success in computer vision and natural language processing, where data samples exhibit explicit spatial or semantic dependencies. However, applying these methods to tabular data is challenging due to the less pronounced dependencies among data samples. In this paper, we address this limitation by introducing SwitchTab, a novel self-supervised method specifically designed to capture latent dependencies in tabular data. SwitchTab leverages an asymmetric encoder-decoder framework to decouple mutual and salient features among data pairs, resulting in more representative embeddings. These embeddings, in turn, contribute to better decision boundaries and lead to improved results in downstream tasks. To validate the effectiveness of SwitchTab, we conduct extensive experiments across various domains involving tabular data. The results showcase superior performance in end-to-end prediction tasks with fine-tuning. Moreover, we demonstrate that pre-trained salient embeddings can be utilized as plug-and-play features to enhance the performance of various traditional classification methods (e.g., Logistic Regression, XGBoost, etc.). Lastly, we highlight the capability of SwitchTab to create explainable representations through visualization of decoupled mutual and salient features in the latent space.

cs.LG↗

arXiv2023

FP8-LM: Training FP8 Large Language Models

In this paper, we explore FP8 low-bit data formats for efficient training of large language models (LLMs). Our key insight is that most variables, such as gradients and optimizer states, in LLM training can employ low-precision data formats without compromising model accuracy and requiring no changes to hyper-parameters. Specifically, we propose a new FP8 automatic mixed-precision framework for training LLMs. This framework offers three levels of FP8 utilization to streamline mixed-precision and distributed parallel training for LLMs. It gradually incorporates 8-bit gradients, optimizer states, and distributed learning in an incremental manner. Experiment results show that, during the training of GPT-175B model on H100 GPU platform, our FP8 mixed-precision training framework not only achieved a remarkable 39% reduction in real memory usage but also ran 75% faster than the widely adopted BF16 framework (i.e., Megatron-LM), surpassing the speed of Nvidia Transformer Engine by 37%. This largely reduces the training costs for large foundation models. Furthermore, our FP8 mixed-precision training methodology is generic. It can be seamlessly applied to other tasks such as LLM instruction tuning and reinforcement learning with human feedback, offering savings in fine-tuning expenses. Our FP8 low-precision training framework is open-sourced at {https://github.com/Azure/MS-AMP}{aka.ms/MS.AMP}.

cs.LG↗

arXiv2023

External Knowledge Augmented Polyphone Disambiguation Using Large Language Model

One of the key issues in Mandarin Chinese text-to-speech (TTS) systems is polyphone disambiguation when doing grapheme-to-phoneme (G2P) conversion. In this paper, we introduce a novel method to solve the problem as a generation task. Following the trending research of large language models (LLM) and prompt learning, the proposed method consists of three modules. Retrieval module incorporates external knowledge which is a multi-level semantic dictionary of Chinese polyphonic characters to format the sentence into a prompt. Generation module adopts the decoder-only Transformer architecture to induce the target text. Postprocess module corrects the generated text into a valid result if needed. Experimental results show that our method outperforms the existing methods on a public dataset called CPP. We also empirically study the impacts of different templates of the prompt, different sizes of training data, and whether to incorporate external knowledge.

cs.CL↗

arXiv2023

Relation-Aware Question Answering for Heterogeneous Knowledge Graphs

Multi-hop Knowledge Base Question Answering(KBQA) aims to find the answer entity in a knowledge graph (KG), which requires multiple steps of reasoning. Existing retrieval-based approaches solve this task by concentrating on the specific relation at different hops and predicting the intermediate entity within the reasoning path. During the reasoning process of these methods, the representation of relations are fixed but the initial relation representation may not be optimal. We claim they fail to utilize information from head-tail entities and the semantic connection between relations to enhance the current relation representation, which undermines the ability to capture information of relations in KGs. To address this issue, we construct a \textbf{dual relation graph} where each node denotes a relation in the original KG (\textbf{primal entity graph}) and edges are constructed between relations sharing same head or tail entities. Then we iteratively do primal entity graph reasoning, dual relation graph information propagation, and interaction between these two graphs. In this way, the interaction between entity and relation is enhanced, and we derive better entity and relation representations. Experiments on two public datasets, WebQSP and CWQ, show that our approach achieves a significant performance gain over the prior state-of-the-art. Our code is available on \url{https://github.com/yanmenxue/RAH-KBQA}.

cs.CL↗

arXiv2023

GHuNeRF: Generalizable Human NeRF from a Monocular Video

In this paper, we tackle the challenging task of learning a generalizable human NeRF model from a monocular video. Although existing generalizable human NeRFs have achieved impressive results, they require muti-view images or videos which might not be always available. On the other hand, some works on free-viewpoint rendering of human from monocular videos cannot be generalized to unseen identities. In view of these limitations, we propose GHuNeRF to learn a generalizable human NeRF model from a monocular video of the human performer. We first introduce a visibility-aware aggregation scheme to compute vertex-wise features, which is used to construct a 3D feature volume. The feature volume can only represent the overall geometry of the human performer with insufficient accuracy due to the limited resolution. To solve this, we further enhance the volume feature with temporally aligned point-wise features using an attention mechanism. Finally, the enhanced feature is used for predicting density and color for each sampled point. A surface-guided sampling strategy is also adopted to improve the efficiency for both training and inference. We validate our approach on the widely-used ZJU-MoCap dataset, where we achieve comparable performance with existing multi-view video based approaches. We also test on the monocular People-Snapshot dataset and achieve better performance than existing works when only monocular video is used. Our code is available at the project website.

cs.CV↗

arXiv2023

How to Determine the Most Powerful Pre-trained Language Model without Brute Force Fine-tuning? An Empirical Survey

Transferability estimation has been attached to great attention in the computer vision fields. Researchers try to estimate with low computational cost the performance of a model when transferred from a source task to a given target task. Considering the effectiveness of such estimations, the communities of natural language processing also began to study similar problems for the selection of pre-trained language models. However, there is a lack of a comprehensive comparison between these estimation methods yet. Also, the differences between vision and language scenarios make it doubtful whether previous conclusions can be established across fields. In this paper, we first conduct a thorough survey of existing transferability estimation methods being able to find the most suitable model, then we conduct a detailed empirical study for the surveyed methods based on the GLUE benchmark. From qualitative and quantitative analyses, we demonstrate the strengths and weaknesses of existing methods and show that H-Score generally performs well with superiorities in effectiveness and efficiency. We also outline the difficulties of consideration of training details, applicability to text generation, and consistency to certain metrics which shed light on future directions.

cs.CL↗

arXiv2023

Noise Adaptor in Spiking Neural Networks

Recent strides in low-latency spiking neural network (SNN) algorithms have drawn significant interest, particularly due to their event-driven computing nature and fast inference capability. One of the most efficient ways to construct a low-latency SNN is by converting a pre-trained, low-bit artificial neural network (ANN) into an SNN. However, this conversion process faces two main challenges: First, converting SNNs from low-bit ANNs can lead to ``occasional noise" -- the phenomenon where occasional spikes are generated in spiking neurons where they should not be -- during inference, which significantly lowers SNN accuracy. Second, although low-latency SNNs initially show fast improvements in accuracy with time steps, these accuracy growths soon plateau, resulting in their peak accuracy lagging behind both full-precision ANNs and traditional ``long-latency SNNs'' that prioritize precision over speed. In response to these two challenges, this paper introduces a novel technique named ``noise adaptor.'' Noise adaptor can model occasional noise during training and implicitly optimize SNN accuracy, particularly at high simulation times $T$. Our research utilizes the ResNet model for a comprehensive analysis of the impact of the noise adaptor on low-latency SNNs. The results demonstrate that our method outperforms the previously reported quant-ANN-to-SNN conversion technique. We achieved an accuracy of 95.95\% within 4 time steps on CIFAR-10 using ResNet-18, and an accuracy of 74.37\% within 64 time steps on ImageNet using ResNet-50. Remarkably, these results were obtained without resorting to any noise correction methods during SNN inference, such as negative spikes or two-stage SNN simulations. Our approach significantly boosts the peak accuracy of low-latency SNNs, bringing them on par with the accuracy of full-precision ANNs. Code will be open source.

cs.CV↗