arXiv subjects

Zheng Lin

Publications and source records attributed to Zheng Lin.

At least 163 records · Page 9Linked to original sources

arXiv2024

RealEra: Semantic-level Concept Erasure via Neighbor-Concept Mining

The remarkable development of text-to-image generation models has raised notable security concerns, such as the infringement of portrait rights and the generation of inappropriate content. Concept erasure has been proposed to remove the model's knowledge about protected and inappropriate concepts. Although many methods have tried to balance the efficacy (erasing target concepts) and specificity (retaining irrelevant concepts), they can still generate abundant erasure concepts under the steering of semantically related inputs. In this work, we propose RealEra to address this "concept residue" issue. Specifically, we first introduce the mechanism of neighbor-concept mining, digging out the associated concepts by adding random perturbation into the embedding of erasure concept, thus expanding the erasing range and eliminating the generations even through associated concept inputs. Furthermore, to mitigate the negative impact on the generation of irrelevant concepts caused by the expansion of erasure scope, RealEra preserves the specificity through the beyond-concept regularization. This makes irrelevant concepts maintain their corresponding spatial position, thereby preserving their normal generation performance. We also employ the closed-form solution to optimize weights of U-Net for the cross-attention alignment, as well as the prediction noise alignment with the LoRA module. Extensive experiments on multiple benchmarks demonstrate that RealEra outperforms previous concept erasing methods in terms of superior erasing efficacy, specificity, and generality. More details are available on our project page https://realerasing.github.io/RealEra/ .

cs.CV↗

arXiv2024

An Empirical Study of ChatGPT-Related Projects and Their Issues on GitHub

Since the launch of ChatGPT in 2022, an increasing number of ChatGPT-related projects are being published on GitHub, sparking widespread discussions. However, GitHub does not provide a detailed classification of these projects to help users effectively explore interested projects. Additionally, the issues raised by users for these projects cover various aspects, e.g., installation, usage, and updates. It would be valuable to help developers prioritize more urgent issues and improve development efficiency. We retrieved 71,244 projects from GitHub using the keyword `ChatGPT' and selected the top 200 representative projects with the highest numbers of stars as our dataset. By analyzing the project descriptions, we identified three primary categories of ChatGPT-related projects, namely ChatGPT Implementation & Training, ChatGPT Application, ChatGPT Improvement & Extension. Next, we applied a topic modeling technique to 23,609 issues of those projects and identified ten issue topics, e.g., model reply and interaction interface. We further analyzed the popularity, difficulty, and evolution of each issue topic within the three project categories. Our main findings are: 1) The increase in the number of projects within the three categories is closely related to the development of ChatGPT; and 2) There are significant differences in the popularity, difficulty, and evolutionary trends of the issue topics across the three project categories. Based on these findings, we finally provided implications for project developers and platform managers on how to better develop and manage ChatGPT-related projects.

cs.SE↗

arXiv2024

Towards Flexible Evaluation for Generative Visual Question Answering

Throughout rapid development of multimodal large language models, a crucial ingredient is a fair and accurate evaluation of their multimodal comprehension abilities. Although Visual Question Answering (VQA) could serve as a developed test field, limitations of VQA evaluation, like the inflexible pattern of Exact Match, have hindered MLLMs from demonstrating their real capability and discourage rich responses. Therefore, this paper proposes the use of semantics-based evaluators for assessing unconstrained open-ended responses on VQA datasets. As characteristics of VQA have made such evaluation significantly different than the traditional Semantic Textual Similarity (STS) task, to systematically analyze the behaviour and compare the performance of various evaluators including LLM-based ones, we proposes three key properties, i.e., Alignment, Consistency and Generalization, and a corresponding dataset Assessing VQA Evaluators (AVE) to facilitate analysis. In addition, this paper proposes a Semantically Flexible VQA Evaluator (SFVE) with meticulous design based on the unique features of VQA evaluation. Experimental results verify the feasibility of model-based VQA evaluation and effectiveness of the proposed evaluator that surpasses existing semantic evaluators by a large margin. The proposed training scheme generalizes to both the BERT-like encoders and decoder-only LLM.

cs.CV↗

arXiv2024

Interpreting and Mitigating Hallucination in MLLMs through Multi-agent Debate

MLLMs often generate outputs that are inconsistent with the visual content, a challenge known as hallucination. Previous methods focus on determining whether a generated output is hallucinated, without identifying which image region leads to the hallucination or interpreting why such hallucinations occur. In this paper, we argue that hallucination in MLLMs is partially due to a lack of slow-thinking and divergent-thinking in these models. To address this, we propose adopting a self-reflection scheme to promote slow-thinking. Furthermore, we consider eliminating hallucination as a complex reasoning task and propose a multi-agent debate approach to encourage divergent-thinking. Consequently, our approach can not only mitigate hallucinations but also interpret why they occur and detail the specifics of hallucination. In addition, we propose to distinguish creativity from hallucination in the context of MLLMs, and illustrate how to evaluate MLLMs' creativity capability. Extensive experiments on various benchmarks demonstrate that our approach exhibits generalized hallucinations-mitigating performance across several MLLMs.

cs.CV↗

arXiv2024

Disrupting Diffusion: Token-Level Attention Erasure Attack against Diffusion-based Customization

With the development of diffusion-based customization methods like DreamBooth, individuals now have access to train the models that can generate their personalized images. Despite the convenience, malicious users have misused these techniques to create fake images, thereby triggering a privacy security crisis. In light of this, proactive adversarial attacks are proposed to protect users against customization. The adversarial examples are trained to distort the customization model's outputs and thus block the misuse. In this paper, we propose DisDiff (Disrupting Diffusion), a novel adversarial attack method to disrupt the diffusion model outputs. We first delve into the intrinsic image-text relationships, well-known as cross-attention, and empirically find that the subject-identifier token plays an important role in guiding image generation. Thus, we propose the Cross-Attention Erasure module to explicitly "erase" the indicated attention maps and disrupt the text guidance. Besides,we analyze the influence of the sampling process of the diffusion model on Projected Gradient Descent (PGD) attack and introduce a novel Merit Sampling Scheduler to adaptively modulate the perturbation updating amplitude in a step-aware manner. Our DisDiff outperforms the state-of-the-art methods by 12.75% of FDFR scores and 7.25% of ISM scores across two facial benchmarks and two commonly used prompts on average.

cs.CV↗

arXiv2024

Pruning Large Language Models to Intra-module Low-rank Architecture with Transitional Activations

Structured pruning fundamentally reduces computational and memory overheads of large language models (LLMs) and offers a feasible solution for end-side LLM deployment. Structurally pruned models remain dense and high-precision, highly compatible with further tuning and compression. However, as the coarse-grained structured pruning poses large damage to the highly interconnected model, achieving a high compression ratio for scaled-up LLMs remains a challenge. In this paper, we introduce a task-agnostic structured pruning approach coupled with a compact Transformer architecture design. The proposed approach, named TransAct, reduces transitional activations inside multi-head attention (MHA) and multi-layer perceptron (MLP) modules, while preserving the inter-module activations that are sensitive to perturbations. Hence, the LLM is pruned into an intra-module low-rank architecture, significantly reducing weights, KV Cache and attention computation. TransAct is implemented on the LLaMA model and evaluated on downstream benchmarks. Results verify the optimality of our approach at high compression with respect to both efficiency and performance. Further, ablation studies reveal the strength of activation-guided iterative pruning and provide experimental analysis on the redundancy of MHA and MLP modules.

cs.CL↗

arXiv2024

DiffRetouch: Using Diffusion to Retouch on the Shoulder of Experts

Image retouching aims to enhance the visual quality of photos. Considering the different aesthetic preferences of users, the target of retouching is subjective. However, current retouching methods mostly adopt deterministic models, which not only neglects the style diversity in the expert-retouched results and tends to learn an average style during training, but also lacks sample diversity during inference. In this paper, we propose a diffusion-based method, named DiffRetouch. Thanks to the excellent distribution modeling ability of diffusion, our method can capture the complex fine-retouched distribution covering various visual-pleasing styles in the training data. Moreover, four image attributes are made adjustable to provide a user-friendly editing mechanism. By adjusting these attributes in specified ranges, users are allowed to customize preferred styles within the learned fine-retouched distribution. Additionally, the affine bilateral grid and contrastive learning scheme are introduced to handle the problem of texture distortion and control insensitivity respectively. Extensive experiments have demonstrated the superior performance of our method on visually appealing and sample diversity. The code will be made available to the community.

cs.CV↗

arXiv2024

SplitLoRA: A Split Parameter-Efficient Fine-Tuning Framework for Large Language Models

The scalability of large language models (LLMs) in handling high-complexity models and large-scale datasets has led to tremendous successes in pivotal domains. While there is an urgent need to acquire more training data for LLMs, a concerning reality is the depletion of high-quality public datasets within a few years. In view of this, the federated learning (FL) LLM fine-tuning paradigm recently has been proposed to facilitate collaborative LLM fine-tuning on distributed private data, where multiple data owners collaboratively fine-tune a shared LLM without sharing raw data. However, the staggering model size of LLMs imposes heavy computing and communication burdens on clients, posing significant barriers to the democratization of the FL LLM fine-tuning paradigm. To address this issue, split learning (SL) has emerged as a promising solution by offloading the primary training workload to a server via model partitioning while exchanging activation/activation's gradients with smaller data sizes rather than the entire LLM. Unfortunately, research on the SL LLM fine-tuning paradigm is still in its nascent stage. To fill this gap, in this paper, we propose the first SL LLM fine-tuning framework, named SplitLoRA. SplitLoRA is built on the split federated learning (SFL) framework, amalgamating the advantages of parallel training from FL and model splitting from SL and thus greatly enhancing the training efficiency. It is worth noting that SplitLoRA is the inaugural open-source benchmark for SL LLM fine-tuning, providing a foundation for research efforts dedicated to advancing SL LLM fine-tuning. Extensive simulations validate that SplitLoRA achieves target accuracy in significantly less time than state-of-the-art LLM fine-tuning frameworks, demonstrating the superior training performance of SplitLoRA. The project page is available at https://fduinc.github.io/splitlora/.

cs.LG↗

arXiv2024

LEO Satellite Networks Assisted Geo-distributed Data Processing

Nowadays, the increasing deployment of edge clouds globally provides users with low-latency services. However, connecting an edge cloud to a core cloud via optic cables in terrestrial networks poses significant barriers due to the prohibitively expensive building cost of optic cables. Fortunately, emerging Low Earth Orbit (LEO) satellite networks (e.g., Starlink) offer a more cost-effective solution for increasing edge clouds, and hence large volumes of data in edge clouds can be transferred to a core cloud via those networks for time-sensitive big data tasks processing, such as attack detection. However, the state-of-the-art satellite selection algorithms bring poor performance for those processing via our measurements. Therefore, we propose a novel data volume aware satellite selection algorithm, named DVA, to support such big data processing tasks. DVA first takes into account both the data size in edge clouds and satellite capacity to finalize the selection, thereby preventing congestion in the access network and reducing transmitting duration. Extensive simulations validate that DVA has a significantly lower average access network duration than the state-of-the-art satellite selection algorithms in a LEO satellite emulation platform.

cs.NI↗

arXiv2024

Multimodal Table Understanding

Although great progress has been made by previous table understanding methods including recent approaches based on large language models (LLMs), they rely heavily on the premise that given tables must be converted into a certain text sequence (such as Markdown or HTML) to serve as model input. However, it is difficult to access such high-quality textual table representations in some real-world scenarios, and table images are much more accessible. Therefore, how to directly understand tables using intuitive visual information is a crucial and urgent challenge for developing more practical applications. In this paper, we propose a new problem, multimodal table understanding, where the model needs to generate correct responses to various table-related requests based on the given table image. To facilitate both the model training and evaluation, we construct a large-scale dataset named MMTab, which covers a wide spectrum of table images, instructions and tasks. On this basis, we develop Table-LLaVA, a generalist tabular multimodal large language model (MLLM), which significantly outperforms recent open-source MLLM baselines on 23 benchmarks under held-in and held-out settings. The code and data is available at this https://github.com/SpursGoZmy/Table-LLaVA

cs.CL↗

arXiv2024

Think out Loud: Emotion Deducing Explanation in Dialogues

Humans convey emotions through daily dialogues, making emotion understanding a crucial step of affective intelligence. To understand emotions in dialogues, machines are asked to recognize the emotion for an utterance (Emotion Recognition in Dialogues, ERD); based on the emotion, then find causal utterances for the emotion (Emotion Cause Extraction in Dialogues, ECED). The setting of the two tasks requires first ERD and then ECED, ignoring the mutual complement between emotion and cause. To fix this, some new tasks are proposed to extract them simultaneously. Although the current research on these tasks has excellent achievements, simply identifying emotion-related factors by classification modeling lacks realizing the specific thinking process of causes stimulating the emotion in an explainable way. This thinking process especially reflected in the reasoning ability of Large Language Models (LLMs) is under-explored. To this end, we propose a new task "Emotion Deducing Explanation in Dialogues" (EDEN). EDEN recognizes emotion and causes in an explicitly thinking way. That is, models need to generate an explanation text, which first summarizes the causes; analyzes the inner activities of the speakers triggered by the causes using common sense; then guesses the emotion accordingly. To support the study of EDEN, based on the existing resources in ECED, we construct two EDEN datasets by human effort. We further evaluate different models on EDEN and find that LLMs are more competent than conventional PLMs. Besides, EDEN can help LLMs achieve better recognition of emotions and causes, which explores a new research direction of explainable emotion understanding in dialogues.

cs.CL↗

arXiv2024

Light-PEFT: Lightening Parameter-Efficient Fine-Tuning via Early Pruning

Parameter-efficient fine-tuning (PEFT) has emerged as the predominant technique for fine-tuning in the era of large language models. However, existing PEFT methods still have inadequate training efficiency. Firstly, the utilization of large-scale foundation models during the training process is excessively redundant for certain fine-tuning tasks. Secondly, as the model size increases, the growth in trainable parameters of empirically added PEFT modules becomes non-negligible and redundant, leading to inefficiency. To achieve task-specific efficient fine-tuning, we propose the Light-PEFT framework, which includes two methods: Masked Early Pruning of the Foundation Model and Multi-Granularity Early Pruning of PEFT. The Light-PEFT framework allows for the simultaneous estimation of redundant parameters in both the foundation model and PEFT modules during the early stage of training. These parameters can then be pruned for more efficient fine-tuning. We validate our approach on GLUE, SuperGLUE, QA tasks, and various models. With Light-PEFT, parameters of the foundation model can be pruned by up to over 40%, while still controlling trainable parameters to be only 25% of the original PEFT method. Compared to utilizing the PEFT method directly, Light-PEFT achieves training and inference speedup, reduces memory usage, and maintains comparable performance and the plug-and-play feature of PEFT.

cs.CL↗

arXiv2024

SATSense: Multi-Satellite Collaborative Framework for Spectrum Sensing

Low Earth Orbit satellite Internet has recently been deployed, providing worldwide service with non-terrestrial networks. With the large-scale deployment of both non-terrestrial and terrestrial networks, limited spectrum resources will not be allocated enough. Consequently, dynamic spectrum sharing is crucial for their coexistence in the same spectrum, where accurate spectrum sensing is essential. However, spectrum sensing in space is more challenging than in terrestrial networks due to variable channel conditions, making single-satellite sensing unstable. Therefore, we first attempt to design a collaborative sensing scheme utilizing diverse data from multiple satellites. However, it is non-trivial to achieve this collaboration due to heterogeneous channel quality, considerable raw sampling data, and packet loss. To address the above challenges, we first establish connections between the satellites by modeling their sensing data as a graph and devising a graph neural network-based algorithm to achieve effective spectrum sensing. Meanwhile, we establish a joint sub-Nyquist sampling and autoencoder data compression framework to reduce the amount of transmitted sensing data. Finally, we propose a contrastive learning-based mechanism compensates for missing packets. Extensive experiments demonstrate that our proposed strategy can achieve efficient spectrum sensing performance and outperform the conventional deep learning algorithm in spectrum sensing accuracy.

cs.NI↗

arXiv2024

Sums: Sniffing Unknown Multiband Signals under Low Sampling Rates

Due to sophisticated deployments of all kinds of wireless networks (e.g., 5G, Wi-Fi, Bluetooth, LEO satellite, etc.), multiband signals distribute in a large bandwidth (e.g., from 70 MHz to 8 GHz). Consequently, for network monitoring and spectrum sharing applications, a sniffer for extracting physical layer information, such as structure of packet, with low sampling rate (especially, sub-Nyquist sampling) can significantly improve their cost- and energy-efficiency. However, to achieve a multiband signals sniffer is really a challenge. To this end, we propose Sums, a system that can sniff and analyze multiband signals in a blind manner. Our Sums takes advantage of hardware and algorithm co-design, multi-coset sub-Nyquist sampling hardware, and a multi-task deep learning framework. The hardware component breaks the Nyquist rule to sample GHz bandwidth, but only pays for a 50 MSPS sampling rate. Our multi-task learning framework directly tackles the sampling data to perform spectrum sensing, physical layer protocol recognition, and demodulation for deep inspection from multiband signals. Extensive experiments demonstrate that Sums achieves higher accuracy than the state-of-theart baselines in spectrum sensing, modulation classification, and demodulation. As a result, our Sums can help researchers and end-users to diagnose or troubleshoot their problems of wireless infrastructures deployments in practice.

cs.AR↗

arXiv2024

TrimCaching: Parameter-sharing AI Model Caching in Wireless Edge Networks

Next-generation mobile networks are expected to facilitate fast AI model downloading to end users. By caching models on edge servers, mobile networks can deliver models to end users with low latency, resulting in a paradigm called edge model caching. In this paper, we develop a novel model placement scheme, called parameter-sharing model caching (TrimCaching). TrimCaching exploits the key observation that a wide range of AI models, such as convolutional neural networks or large language models, can share a significant proportion of parameter blocks containing reusable knowledge, thereby improving storage efficiency. To this end, we formulate a parameter-sharing model placement problem to maximize the cache hit ratio in multi-edge wireless networks by balancing the fundamental tradeoff between storage efficiency and service latency. We show that the formulated problem is a submodular maximization problem with submodular constraints, for which no polynomial-time approximation algorithm exists. To overcome this challenge, we study an important special case, where a small fixed number of parameter blocks are shared across models, which often holds in practice. In such a case, a polynomial-time algorithm with $\left(1-ε\right)/2$-approximation guarantee is developed. Subsequently, we address the original problem for the general case by developing a greedy algorithm. Simulation results demonstrate that the proposed TrimCaching framework significantly improves the cache hit ratio compared with state-of-the-art content caching without exploiting shared parameters in AI models.

cs.NI↗

arXiv2024

Deep Learning Inference on Heterogeneous Mobile Processors: Potentials and Pitfalls

There is a growing demand to deploy computation-intensive deep learning (DL) models on resource-constrained mobile devices for real-time intelligent applications. Equipped with a variety of processing units such as CPUs, GPUs, and NPUs, the mobile devices hold potential to accelerate DL inference via parallel execution across heterogeneous processors. Various efficient parallel methods have been explored to optimize computation distribution, achieve load balance, and minimize communication cost across processors. Yet their practical effectiveness in the dynamic and diverse real-world mobile environment is less explored. This paper presents a holistic empirical study to assess the capabilities and challenges associated with parallel DL inference on heterogeneous mobile processors. Through carefully designed experiments covering various DL models, mobile software/hardware environments, workload patterns, and resource availability, we identify limitations of existing techniques and highlight opportunities for cross-level optimization.

cs.LG↗

arXiv2024

AdaOper: Energy-efficient and Responsive Concurrent DNN Inference on Mobile Devices

Deep neural network (DNN) has driven extensive applications in mobile technology. However, for long-running mobile apps like voice assistants or video applications on smartphones, energy efficiency is critical for battery-powered devices. The rise of heterogeneous processors in mobile devices today has introduced new challenges for optimizing energy efficiency. Our key insight is that partitioning computations across different processors for parallelism and speedup doesn't necessarily correlate with energy consumption optimization and may even increase it. To address this, we present AdaOper, an energy-efficient concurrent DNN inference system. It optimizes energy efficiency on mobile heterogeneous processors while maintaining responsiveness. AdaOper includes a runtime energy profiler that dynamically adjusts operator partitioning to optimize energy efficiency based on dynamic device conditions. We conduct preliminary experiments, which show that AdaOper reduces energy consumption by 16.88% compared to the existing concurrent method while ensuring real-time performance.

cs.DC↗

arXiv2024

IFViT: Interpretable Fixed-Length Representation for Fingerprint Matching via Vision Transformer

Determining dense feature points on fingerprints used in constructing deep fixed-length representations for accurate matching, particularly at the pixel level, is of significant interest. To explore the interpretability of fingerprint matching, we propose a multi-stage interpretable fingerprint matching network, namely Interpretable Fixed-length Representation for Fingerprint Matching via Vision Transformer (IFViT), which consists of two primary modules. The first module, an interpretable dense registration module, establishes a Vision Transformer (ViT)-based Siamese Network to capture long-range dependencies and the global context in fingerprint pairs. It provides interpretable dense pixel-wise correspondences of feature points for fingerprint alignment and enhances the interpretability in the subsequent matching stage. The second module takes into account both local and global representations of the aligned fingerprint pair to achieve an interpretable fixed-length representation extraction and matching. It employs the ViTs trained in the first module with the additional fully connected layer and retrains them to simultaneously produce the discriminative fixed-length representation and interpretable dense pixel-wise correspondences of feature points. Extensive experimental results on diverse publicly available fingerprint databases demonstrate that the proposed framework not only exhibits superior performance on dense registration and matching but also significantly promotes the interpretability in deep fixed-length representations-based fingerprint matching.

cs.CV↗