arXiv subjects

Liang Chen

Publications and source records attributed to Liang Chen.

At least 253 records · Page 14Linked to original sources

arXiv2024

SGHormer: An Energy-Saving Graph Transformer Driven by Spikes

Graph Transformers (GTs) with powerful representation learning ability make a huge success in wide range of graph tasks. However, the costs behind outstanding performances of GTs are higher energy consumption and computational overhead. The complex structure and quadratic complexity during attention calculation in vanilla transformer seriously hinder its scalability on the large-scale graph data. Though existing methods have made strides in simplifying combinations among blocks or attention-learning paradigm to improve GTs' efficiency, a series of energy-saving solutions originated from biologically plausible structures are rarely taken into consideration when constructing GT framework. To this end, we propose a new spiking-based graph transformer (SGHormer). It turns full-precision embeddings into sparse and binarized spikes to reduce memory and computational costs. The spiking graph self-attention and spiking rectify blocks in SGHormer explicitly capture global structure information and recover the expressive power of spiking embeddings, respectively. In experiments, SGHormer achieves comparable performances to other full-precision GTs with extremely low computational energy consumption. The results show that SGHomer makes a remarkable progress in the field of low-energy GTs.

cs.NE↗

arXiv2024

Measurements of All-Particle Energy Spectrum and Mean Logarithmic Mass of Cosmic Rays from 0.3 to 30 PeV with LHAASO-KM2A

We present the measurements of all-particle energy spectrum and mean logarithmic mass of cosmic rays in the energy range of 0.3-30 PeV using data collected from LHAASO-KM2A between September 2021 and December 2022, which is based on a nearly composition-independent energy reconstruction method, achieving unprecedented accuracy. Our analysis reveals the position of the knee at $3.67 \pm 0.05 \pm 0.15$ PeV. Below the knee, the spectral index is found to be -$2.7413 \pm 0.0004 \pm 0.0050$, while above the knee, it is -$3.128 \pm 0.005 \pm 0.027$, with the sharpness of the transition measured with a statistical error of 2%. The mean logarithmic mass of cosmic rays is almost heavier than helium in the whole measured energy range. It decreases from 1.7 at 0.3 PeV to 1.3 at 3 PeV, representing a 24% decline following a power law with an index of -$0.1200 \pm 0.0003 \pm 0.0341$. This is equivalent to an increase in abundance of light components. Above the knee, the mean logarithmic mass exhibits a power law trend towards heavier components, which is reversal to the behavior observed in the all-particle energy spectrum. Additionally, the knee position and the change in power-law index are approximately the same. These findings suggest that the knee observed in the all-particle spectrum corresponds to the knee of the light component, rather than the medium-heavy components.

astro-ph.HE↗

arXiv2024

Self-Guard: Empower the LLM to Safeguard Itself

The jailbreak attack can bypass the safety measures of a Large Language Model (LLM), generating harmful content. This misuse of LLM has led to negative societal consequences. Currently, there are two main approaches to address jailbreak attacks: safety training and safeguards. Safety training focuses on further training LLM to enhance its safety. On the other hand, safeguards involve implementing external models or filters to prevent harmful outputs. However, safety training has constraints in its ability to adapt to new attack types and often leads to a drop in model performance. Safeguards have proven to be of limited help. To tackle these issues, we propose a novel approach called Self-Guard, which combines the strengths of both safety methods. Self-Guard includes two stages. In the first stage, we enhance the model's ability to assess harmful content, and in the second stage, we instruct the model to consistently perform harmful content detection on its own responses. The experiment has demonstrated that Self-Guard is robust against jailbreak attacks. In the bad case analysis, we find that LLM occasionally provides harmless responses to harmful queries. Additionally, we evaluated the general capabilities of the LLM before and after safety training, providing evidence that Self-Guard does not result in the LLM's performance degradation. In sensitivity tests, Self-Guard not only avoids inducing over-sensitivity in LLM but also can even mitigate this issue.

cs.CL↗

arXiv2024

MMICL: Empowering Vision-language Model with Multi-Modal In-Context Learning

Since the resurgence of deep learning, vision-language models (VLMs) enhanced by large language models (LLMs) have grown exponentially in popularity. However, while LLMs can utilize extensive background knowledge and task information with in-context learning, most VLMs still struggle with understanding complex multi-modal prompts with multiple images, making VLMs less effective in downstream vision-language tasks. In this paper, we address the limitation above by 1) introducing vision-language Model with Multi-Modal In-Context Learning(MMICL), a new approach to allow the VLM to deal with multi-modal inputs efficiently; 2) proposing a novel context scheme to augment the in-context learning ability of the VLM; 3) constructing the Multi-modal In-Context Learning (MIC) dataset, designed to enhance the VLM's ability to understand complex multi-modal prompts. Our experiments confirm that MMICL achieves new state-of-the-art zero-shot performance on a wide range of general vision-language tasks, especially for complex benchmarks, including MME and MMBench. Our analysis demonstrates that MMICL effectively tackles the challenge of complex multi-modal prompt understanding and emerges the impressive ICL ability. Furthermore, we observe that MMICL successfully alleviates language bias in VLMs, a common issue for VLMs that often leads to hallucination when faced with extensive textual context. Our code, dataset, dataset tool, and model are available at https://github.com/PKUnlp-icler/MIC

cs.CL↗

arXiv2024

A Causal Inspired Early-Branching Structure for Domain Generalization

Learning domain-invariant semantic representations is crucial for achieving domain generalization (DG), where a model is required to perform well on unseen target domains. One critical challenge is that standard training often results in entangled semantic and domain-specific features. Previous works suggest formulating the problem from a causal perspective and solving the entanglement problem by enforcing marginal independence between the causal (\ie semantic) and non-causal (\ie domain-specific) features. Despite its simplicity, the basic marginal independent-based idea alone may be insufficient to identify the causal feature. By d-separation, we observe that the causal feature can be further characterized by being independent of the domain conditioned on the object, and we propose the following two strategies as complements for the basic framework. First, the observation implicitly implies that for the same object, the causal feature should not be associated with the non-causal feature, revealing that the common practice of obtaining the two features with a shared base feature extractor and two lightweight prediction heads might be inappropriate. To meet the constraint, we propose a simple early-branching structure, where the causal and non-causal feature obtaining branches share the first few blocks while diverging thereafter, for better structure design; Second, the observation implies that the causal feature remains invariant across different domains for the same object. To this end, we suggest that augmentation should be incorporated into the framework to better characterize the causal feature, and we further suggest an effective random domain sampling scheme to fulfill the task. Theoretical and experimental results show that the two strategies are beneficial for the basic marginal independent-based framework. Code is available at \url{https://github.com/liangchen527/CausEB}.

cs.CV↗

arXiv2024

PeriodicLoRA: Breaking the Low-Rank Bottleneck in LoRA Optimization

Supervised fine-tuning is the most common method to adapt large language models (LLMs) to downstream tasks, but full fine-tuning LLMs requires massive computational resources. Recently, parameter-efficient fine-tuning (PEFT) methods have been widely studied due to its cost-effectiveness. LoRA is one of the most widely used methods, which assumes that the optimization process is essentially low-dimensional. Although LoRA fine-tuning is effective, there is still a performance gap compared to full fine-tuning, since its weight update is limited to low-rank matrices. In order to break the low-rank bottleneck in LoRA Optimization, we propose PeriodicLoRA (PLoRA), which accumulates low-rank update matrices multiple times to achieve a higher update rank. PLoRA has multiple training stages. During each stage, we still update only the LoRA weights. However, at the end of each stage, we unload the LoRA weights into the backbone parameters and then reinitialize the LoRA states. Experimental results show that PLoRA has stronger learning ability, approximately 1.8 times that of LoRA's learning ability at most, but it does not increase memory usage. Further, we introduce a momentum-based unloading strategy for PLoRA to mitigate the training instability.

cs.CL↗

arXiv2024

Oversmoothing: A Nightmare for Graph Contrastive Learning?

Oversmoothing is a common phenomenon observed in graph neural networks (GNNs), in which an increase in the network depth leads to a deterioration in their performance. Graph contrastive learning (GCL) is emerging as a promising way of leveraging vast unlabeled graph data. As a marriage between GNNs and contrastive learning, it remains unclear whether GCL inherits the same oversmoothing defect from GNNs. This work undertakes a fundamental analysis of GCL from the perspective of oversmoothing on the first hand. We demonstrate empirically that increasing network depth in GCL also leads to oversmoothing in their deep representations, and surprisingly, the shallow ones. We refer to this phenomenon in GCL as `long-range starvation', wherein lower layers in deep networks suffer from degradation due to the lack of sufficient guidance from supervision. Based on our findings, we present BlockGCL, a remarkably simple yet effective blockwise training framework to prevent GCL from notorious oversmoothing. Without bells and whistles, BlockGCL consistently improves robustness and stability for well-established GCL methods with increasing numbers of layers on several real-world graph benchmarks.

cs.LG↗

arXiv2024

PCA-Bench: Evaluating Multimodal Large Language Models in Perception-Cognition-Action Chain

We present PCA-Bench, a multimodal decision-making benchmark for evaluating the integrated capabilities of Multimodal Large Language Models (MLLMs). Departing from previous benchmarks focusing on simplistic tasks and individual model capability, PCA-Bench introduces three complex scenarios: autonomous driving, domestic robotics, and open-world games. Given task instructions and diverse contexts, the model is required to seamlessly integrate multiple capabilities of Perception, Cognition, and Action in a reasoning chain to make accurate decisions. Moreover, PCA-Bench features error localization capabilities, scrutinizing model inaccuracies in areas such as perception, knowledge, or reasoning. This enhances the reliability of deploying MLLMs. To balance accuracy and efficiency in evaluation, we propose PCA-Eval, an automatic evaluation protocol, and assess 10 prevalent MLLMs. The results reveal significant performance disparities between open-source models and powerful proprietary models like GPT-4 Vision. To address this, we introduce Embodied-Instruction-Evolution (EIE), an automatic framework for synthesizing instruction tuning examples in multimodal embodied environments. EIE generates 7,510 training examples in PCA-Bench and enhances the performance of open-source MLLMs, occasionally surpassing GPT-4 Vision (+3\% in decision accuracy), thereby validating the effectiveness of EIE. Our findings suggest that robust MLLMs like GPT4-Vision show promise for decision-making in embodied agents, opening new avenues for MLLM research.

cs.CL↗

arXiv2024

A Graph is Worth 1-bit Spikes: When Graph Contrastive Learning Meets Spiking Neural Networks

While contrastive self-supervised learning has become the de-facto learning paradigm for graph neural networks, the pursuit of higher task accuracy requires a larger hidden dimensionality to learn informative and discriminative full-precision representations, raising concerns about computation, memory footprint, and energy consumption burden (largely overlooked) for real-world applications. This work explores a promising direction for graph contrastive learning (GCL) with spiking neural networks (SNNs), which leverage sparse and binary characteristics to learn more biologically plausible and compact representations. We propose SpikeGCL, a novel GCL framework to learn binarized 1-bit representations for graphs, making balanced trade-offs between efficiency and performance. We provide theoretical guarantees to demonstrate that SpikeGCL has comparable expressiveness with its full-precision counterparts. Experimental results demonstrate that, with nearly 32x representation storage compression, SpikeGCL is either comparable to or outperforms many fancy state-of-the-art supervised and self-supervised methods across several graph benchmarks.

cs.NE↗

arXiv2024

Evolution of magnetic field of the Quasar 1604+159 at pc scale

We have analyzed the total intensity, spectral index, linear polarization, and RM distributions at pc scale for the quasar 1604+159. The source was observed in 2002 and 2020 with the VLBA. Combining the MOJAVE results, we studied the evolution of the magnetic field. We detected a core-jet structure. The jet extends to a distance of ~25 mas. The jet shape varies slightly with time. We divided the source structure into the central region and the jet region. In the jet region, we find the polarized emission varies with time. The flatter spectral index values and EVPA direction indicate the possible existence of shocks, contributing to the variation. In the central region, the derived core shift index k_r values indicate that the core in 2002 is close to the equipartition case while deviating from it in 2020. The measured magnetic field strength in 2020 is two orders of magnitude lower than that in 2002. We detected transverse RM gradients, evidence of a helical magnetic field, in the core. At 15 GHz, in the place close to the jet base, the polarization direction changes significantly with time from perpendicular to parallel to the jet direction. The evolution of RM and magnetic field structure are potential reasons for the observed polarization change. The core |RM| in 2020 increases with frequency following a power law with index a = 2.7, suggesting a fast electron density fall-off in the medium with distance from the jet base.

astro-ph.HE↗

arXiv2024

Low power consumption grating magneto-optical trap based on planar elements

The grating-based magneto-optical trap (GMOT) is a promising approach for miniaturizing cold-atom systems. However, the power consumption of a GMOT system dominates its feasibility in practical applications. In this study, we demonstrated a GMOT system based on planar elements that can operate with low power consumption. A high-diffraction-efficiency grating chip was used to cool atoms with a single incident beam. A planar coil chip was designed and fabricated with a low power consumption nested architecture. The grating and coil chips were adapted to a passive pump vacuum chamber, and up to 106 87Rb atoms were trapped. These elements effectively reduce the power consumption of the GMOT and have great potential for applications in practical cold-atom-based devices.

physics.atom-ph↗

arXiv2024

Room-temperature intrinsic ferromagnetism of two-dimensional Na2Cl crystals originated by s- and p-orbitals

Ferromagnetism, as one of the most valuable properties of materials, has attracted sustained and widespread interest in basic and applied research from ancient compasses to modern electronic devices. Traditionally, intrinsic ferromagnetism has been attributed to the permanent magnetic moment induced by partially filled d- or f-orbitals. However, the development of ferromagnetic materials has been limited by this electronic structure convention. Thus, the identification of additional materials that are not constrained by this conventional rule but also exhibit intrinsic ferromagnetism is highly expected and may impact all the fields based on ferromagnetism. Here, we report the direct observation of room-temperature intrinsic ferromagnetism in two-dimensional (2D) Na2Cl crystals, in which there are only partially filled s- and p-orbitals rather than d- or f-orbitals, using the superconducting quantum interference device (SQUID) and magnetic force microscope (MFM). These Na2Cl crystals formed in reduced graphene oxide (rGO) membranes have an unconventional stoichiometric structure leading to unique electron and spin distributions. And the structure of these 2D Na2Cl crystals, including the Na and Cl sites, is characterized in situ for the first time and directly observed by cryo-electron microscopy (cryo-EM) based on the observed difference in contrast between Na stacked with Cl and single Na. These findings break the conventional rule of intrinsic ferromagnetism and provide new insights into the design of novel magnetic and electronic devices and transistors with a size down to the atomic scale.

cond-mat.mtrl-sci↗

arXiv2024

Ultrasensitive piezoelectric sensor based on two-dimensional Na2Cl crystals with periodic atom vacancies

Pursuing ultrasensitivity of pressure sensors has been a long-standing goal. Here, we report a piezoelectric sensor that exhibits supreme pressure-sensing performance, including a peak sensitivity up to 3.5*10^6 kPa^-1 in the pressure range of 1-100 mPa and a detection limit of less than 1 mPa, superior to the current state-of-the-art pressure sensors. These properties are attributed to the high percentage of periodic atom vacancies in the two-dimensional Na2Cl crystals formed within multilayered graphene oxide membrane in the sensor, which provides giant polarization with high stability. The sensor can even clearly detect the airflow fluctuations surrounding a flapping butterfly, which have long been the elusive tiny signals in the famous "butterfly effect". The finding represents a step towards next-generation pressure sensors for various precision applications.

physics.app-ph↗

arXiv2024

Treatment-Aware Hyperbolic Representation Learning for Causal Effect Estimation with Social Networks

Estimating the individual treatment effect (ITE) from observational data is a crucial research topic that holds significant value across multiple domains. How to identify hidden confounders poses a key challenge in ITE estimation. Recent studies have incorporated the structural information of social networks to tackle this challenge, achieving notable advancements. However, these methods utilize graph neural networks to learn the representation of hidden confounders in Euclidean space, disregarding two critical issues: (1) the social networks often exhibit a scalefree structure, while Euclidean embeddings suffer from high distortion when used to embed such graphs, and (2) each ego-centric network within a social network manifests a treatment-related characteristic, implying significant patterns of hidden confounders. To address these issues, we propose a novel method called Treatment-Aware Hyperbolic Representation Learning (TAHyper). Firstly, TAHyper employs the hyperbolic space to encode the social networks, thereby effectively reducing the distortion of confounder representation caused by Euclidean embeddings. Secondly, we design a treatment-aware relationship identification module that enhances the representation of hidden confounders by identifying whether an individual and her neighbors receive the same treatment. Extensive experiments on two benchmark datasets are conducted to demonstrate the superiority of our method.

cs.LG↗

arXiv2024

CodeFuse-13B: A Pretrained Multi-lingual Code Large Language Model

Code Large Language Models (Code LLMs) have gained significant attention in the industry due to their wide applications in the full lifecycle of software engineering. However, the effectiveness of existing models in understanding non-English inputs for multi-lingual code-related tasks is still far from well studied. This paper introduces CodeFuse-13B, an open-sourced pre-trained code LLM. It is specifically designed for code-related tasks with both English and Chinese prompts and supports over 40 programming languages. CodeFuse achieves its effectiveness by utilizing a high quality pre-training dataset that is carefully filtered by program analyzers and optimized during the training process. Extensive experiments are conducted using real-world usage scenarios, the industry-standard benchmark HumanEval-x, and the specially designed CodeFuseEval for Chinese prompts. To assess the effectiveness of CodeFuse, we actively collected valuable human feedback from the AntGroup's software development process where CodeFuse has been successfully deployed. The results demonstrate that CodeFuse-13B achieves a HumanEval pass@1 score of 37.10%, positioning it as one of the top multi-lingual code LLMs with similar parameter sizes. In practical scenarios, such as code generation, code translation, code comments, and testcase generation, CodeFuse performs better than other models when confronted with Chinese prompts.

cs.SE↗

arXiv2024

Expected Transaction Value Optimization for Precise Marketing in FinTech Platforms

FinTech platforms facilitated by digital payments are watching growth rapidly, which enable the distribution of mutual funds personalized to individual investors via mobile Apps. As the important intermediation of financial products investment, these platforms distribute thousands of mutual funds obtaining impressions under guaranteed delivery (GD) strategy required by fund companies. Driven by the profit from fund purchases of users, the platform aims to maximize each transaction amount of customers by promoting mutual funds to these investors who will be interested in. Different from the conversions in traditional advertising or e-commerce recommendations, the investment amount in each purchase varies greatly even for the same financial product, which provides a significant challenge for the promotion recommendation of mutual funds. In addition to predicting the click-through rate (CTR) or the conversion rate (CVR) as in traditional recommendations, it is essential for FinTech platforms to estimate the customers' purchase amount for each delivered fund and achieve an effective allocation of impressions based on the predicted results to optimize the total expected transaction value (ETV). In this paper, we propose an ETV optimized customer allocation framework (EOCA) that aims to maximize the total ETV of recommended funds, under the constraints of GD dealt with fund companies. To the best of our knowledge, it's the first attempt to solve the GD problem for financial product promotions based on customer purchase amount prediction. We conduct extensive experiments on large scale real-world datasets and online tests based on LiCaiTong, Tencent wealth management platform, to demonstrate the effectiveness of our proposed EOCA framework.

cs.IR↗

arXiv2023

Silkie: Preference Distillation for Large Visual Language Models

This paper explores preference distillation for large vision language models (LVLMs), improving their ability to generate helpful and faithful responses anchoring the visual context. We first build a vision-language feedback (VLFeedback) dataset utilizing AI annotation. Specifically, responses are generated by models sampled from 12 LVLMs, conditioned on multi-modal instructions sourced from various datasets. We adopt GPT-4V to assess the generated outputs regarding helpfulness, visual faithfulness, and ethical considerations. Furthermore, the preference supervision is distilled into Qwen-VL-Chat through the direct preference optimization (DPO) method. The resulting model Silkie, achieves 6.9% and 9.5% relative improvement on the MME benchmark regarding the perception and cognition capabilities, respectively. Silkie also demonstrates reduced hallucination by setting a new state-of-the-art score of 3.02 on the MMHal-Bench benchmark. Further analysis shows that DPO with our VLFeedback dataset mainly boosts the fine-grained perception and complex cognition abilities of LVLMs, leading to more comprehensive improvements compared to human-annotated preference datasets.

cs.CV↗

arXiv2023

VQCNIR: Clearer Night Image Restoration with Vector-Quantized Codebook

Night photography often struggles with challenges like low light and blurring, stemming from dark environments and prolonged exposures. Current methods either disregard priors and directly fitting end-to-end networks, leading to inconsistent illumination, or rely on unreliable handcrafted priors to constrain the network, thereby bringing the greater error to the final result. We believe in the strength of data-driven high-quality priors and strive to offer a reliable and consistent prior, circumventing the restrictions of manual priors. In this paper, we propose Clearer Night Image Restoration with Vector-Quantized Codebook (VQCNIR) to achieve remarkable and consistent restoration outcomes on real-world and synthetic benchmarks. To ensure the faithful restoration of details and illumination, we propose the incorporation of two essential modules: the Adaptive Illumination Enhancement Module (AIEM) and the Deformable Bi-directional Cross-Attention (DBCA) module. The AIEM leverages the inter-channel correlation of features to dynamically maintain illumination consistency between degraded features and high-quality codebook features. Meanwhile, the DBCA module effectively integrates texture and structural information through bi-directional cross-attention and deformable convolution, resulting in enhanced fine-grained detail and structural fidelity across parallel decoders. Extensive experiments validate the remarkable benefits of VQCNIR in enhancing image quality under low-light conditions, showcasing its state-of-the-art performance on both synthetic and real-world datasets. The code is available at https://github.com/AlexZou14/VQCNIR.

cs.CV↗