Verifiable Format Control for Large Language Model Generations

Recent Large Language Models (LLMs) have demonstrated satisfying general instruction following ability. However, small LLMs with about 7B parameters still struggle fine-grained format following (e.g., JSON format), which seriously hinder the advancements of their applications. Most existing methods focus on benchmarking general instruction following while overlook how to improve the specific format following ability for small LLMs. Besides, these methods often rely on evaluations based on advanced LLMs (e.g., GPT-4), which can introduce the intrinsic bias of LLMs and be costly due to the API calls. In this paper, we first curate a fully verifiable format following dataset VFF. In contrast to existing works often adopting external LLMs for instruction-following validations, every sample of VFF can be easily validated with a Python function. Further, we propose to leverage this verifiable feature to synthesize massive data for progressively training small LLMs, in order to improve their format following abilities. Experimental results highlight the prevalent limitations in the format following capabilities of 7B level open-source LLMs and demonstrate the effectiveness of our method in enhancing this essential ability.

DOTA-ME-CS: Daily Oriented Text Audio-Mandarin English-Code Switching Dataset

Code-switching, the alternation between two or more languages within communication, poses great challenges for Automatic Speech Recognition (ASR) systems. Existing models and datasets are limited in their ability to effectively handle these challenges. To address this gap and foster progress in code-switching ASR research, we introduce the DOTA-ME-CS: Daily oriented text audio Mandarin-English code-switching dataset, which consists of 18.54 hours of audio data, including 9,300 recordings from 34 participants. To enhance the dataset's diversity, we apply artificial intelligence (AI) techniques such as AI timbre synthesis, speed variation, and noise addition, thereby increasing the complexity and scalability of the task. The dataset is carefully curated to ensure both diversity and quality, providing a robust resource for researchers addressing the intricacies of bilingual speech recognition with detailed data analysis. We further demonstrate the dataset's potential in future research. The DOTA-ME-CS dataset, along with accompanying code, will be made publicly available.

TrustRAG: Enhancing Robustness and Trustworthiness in RAG

Retrieval-Augmented Generation (RAG) systems enhance large language models (LLMs) by integrating external knowledge sources, enabling more accurate and contextually relevant responses tailored to user queries. However, these systems remain vulnerable to corpus poisoning attacks that can significantly degrade LLM performance through the injection of malicious content. To address these challenges, we propose TrustRAG, a robust framework that systematically filters compromised and irrelevant content before it reaches the language model. Our approach implements a two-stage defense mechanism: first, it employs K-means clustering to identify potential attack patterns in retrieved documents based on their semantic embeddings, effectively isolating suspicious content. Second, it leverages cosine similarity and ROUGE metrics to detect malicious documents while resolving discrepancies between the model's internal knowledge and external information through a self-assessment process. TrustRAG functions as a plug-and-play, training-free module that integrates seamlessly with any language model, whether open or closed-source, maintaining high contextual relevance while strengthening defenses against attacks. Through extensive experimental validation, we demonstrate that TrustRAG delivers substantial improvements in retrieval accuracy, efficiency, and attack resistance compared to existing approaches across multiple model architectures and datasets. We have made TrustRAG available as open-source software at \url{https://github.com/HuichiZhou/TrustRAG}.

Can Large Language Models Improve the Adversarial Robustness of Graph Neural Networks?

Graph neural networks (GNNs) are vulnerable to adversarial perturbations, especially for topology attacks, and many methods that improve the robustness of GNNs have received considerable attention. Recently, we have witnessed the significant success of large language models (LLMs), leading many to explore the great potential of LLMs on GNNs. However, they mainly focus on improving the performance of GNNs by utilizing LLMs to enhance the node features. Therefore, we ask: Will the robustness of GNNs also be enhanced with the powerful understanding and inference capabilities of LLMs? By presenting the empirical results, we find that despite that LLMs can improve the robustness of GNNs, there is still an average decrease of 23.1% in accuracy, implying that the GNNs remain extremely vulnerable against topology attack. Therefore, another question is how to extend the capabilities of LLMs on graph adversarial robustness. In this paper, we propose an LLM-based robust graph structure inference framework, LLM4RGNN, which distills the inference capabilities of GPT-4 into a local LLM for identifying malicious edges and an LM-based edge predictor for finding missing important edges, so as to recover a robust graph structure. Extensive experiments demonstrate that LLM4RGNN consistently improves the robustness across various GNNs. Even in some cases where the perturbation ratio increases to 40%, the accuracy of GNNs is still better than that on the clean graph.

Beyond the Hype: A dispassionate look at vision-language models in medical scenario

Recent advancements in Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across diverse tasks, garnering significant attention in AI communities. However, their performance and reliability in specialized domains such as medicine remain insufficiently assessed. In particular, most assessments over-concentrate in evaluating VLMs based on simple Visual Question Answering (VQA) on multi-modality data, while ignoring the in-depth characteristic of LVLMs. In this study, we introduce RadVUQA, a novel Radiological Visual Understanding and Question Answering benchmark, to comprehensively evaluate existing LVLMs. RadVUQA mainly validates LVLMs across five dimensions: 1) Anatomical understanding, assessing the models' ability to visually identify biological structures; 2) Multimodal comprehension, which involves the capability of interpreting linguistic and visual instructions to produce desired outcomes; 3) Quantitative and spatial reasoning, evaluating the models' spatial awareness and proficiency in combining quantitative analysis with visual and linguistic information; 4) Physiological knowledge, measuring the models' capability to comprehend functions and mechanisms of organs and systems; and 5) Robustness, which assesses the models' capabilities against unharmonised and synthetic data. The results indicate that both generalized LVLMs and medical-specific LVLMs have critical deficiencies with weak multimodal comprehension and quantitative reasoning capabilities. Our findings reveal the large gap between existing LVLMs and clinicians, highlighting the urgent need for more robust and intelligent LVLMs. The code and dataset will be available after the acceptance of this paper.

GUI-WORLD: A Dataset for GUI-oriented Multimodal LLM-based Agents

Recently, Multimodal Large Language Models (MLLMs) have been used as agents to control keyboard and mouse inputs by directly perceiving the Graphical User Interface (GUI) and generating corresponding code. However, current agents primarily exhibit excellent understanding capabilities in static environments and are predominantly applied in relatively simple domains, such as Web or mobile interfaces. We argue that a robust GUI agent should be capable of perceiving temporal information on the GUI, including dynamic Web content and multi-step tasks. Additionally, it should possess a comprehensive understanding of various GUI scenarios, including desktop software and multi-window interactions. To this end, this paper introduces a new dataset, termed GUI-World, which features meticulously crafted Human-MLLM annotations, extensively covering six GUI scenarios and eight types of GUI-oriented questions in three formats. We evaluate the capabilities of current state-of-the-art MLLMs, including ImageLLMs and VideoLLMs, in understanding various types of GUI content, especially dynamic and sequential content. Our findings reveal that ImageLLMs struggle with dynamic GUI content without manually annotated keyframes or operation history. On the other hand, VideoLLMs fall short in all GUI-oriented tasks given the sparse GUI video dataset. Based on GUI-World, we take the initial step of leveraging a fine-tuned VideoLLM as a GUI agent, demonstrating an improved understanding of various GUI tasks. However, due to the limitations in the performance of base LLMs, we conclude that using VideoLLMs as GUI agents remains a significant challenge. We believe our work provides valuable insights for future research in dynamic GUI content understanding. The code and dataset are publicly available at our project homepage: https://gui-world.github.io/.

MPAT: Building Robust Deep Neural Networks against Textual Adversarial Attacks

Deep neural networks have been proven to be vulnerable to adversarial examples and various methods have been proposed to defend against adversarial attacks for natural language processing tasks. However, previous defense methods have limitations in maintaining effective defense while ensuring the performance of the original task. In this paper, we propose a malicious perturbation based adversarial training method (MPAT) for building robust deep neural networks against textual adversarial attacks. Specifically, we construct a multi-level malicious example generation strategy to generate adversarial examples with malicious perturbations, which are used instead of original inputs for model training. Additionally, we employ a novel training objective function to ensure achieving the defense goal without compromising the performance on the original task. We conduct comprehensive experiments to evaluate our defense method by attacking five victim models on three benchmark datasets. The result demonstrates that our method is more effective against malicious adversarial attacks compared with previous defense methods while maintaining or further improving the performance on the original task.

MLLM-as-a-Judge: Assessing Multimodal LLM-as-a-Judge with Vision-Language Benchmark

Multimodal Large Language Models (MLLMs) have gained significant attention recently, showing remarkable potential in artificial general intelligence. However, assessing the utility of MLLMs presents considerable challenges, primarily due to the absence multimodal benchmarks that align with human preferences. Inspired by LLM-as-a-Judge in LLMs, this paper introduces a novel benchmark, termed MLLM-as-a-Judge, to assess the ability of MLLMs in assisting judges including three distinct tasks: Scoring Evaluation, Pair Comparison, and Batch Ranking. Our study reveals that, while MLLMs demonstrate remarkable human-like discernment in Pair Comparisons, there is a significant divergence from human preferences in Scoring Evaluation and Batch Ranking tasks. Furthermore, MLLMs still face challenges in judgment, including diverse biases, hallucinatory responses, and inconsistencies, even for advanced models such as GPT-4V. These findings emphasize the pressing need for enhancements and further research efforts regarding MLLMs as fully reliable evaluators. Code and dataset are available at https://github.com/Dongping-Chen/MLLM-as-a-Judge.