The rapid evolution of Multimodal Large Language Models (MLLMs) has brought substantial advancements in artificial intelligence, significantly enhancing the capability to understand and generate multimodal content. While prior studies have largely concentrated on model architectures and training methodologies, a thorough analysis of the benchmarks used for evaluating these models remains underexplored. This survey addresses this gap by systematically reviewing 211 benchmarks that assess MLLMs across four core domains: understanding, reasoning, generation, and application. We provide a detailed analysis of task designs, evaluation metrics, and dataset constructions, across diverse modalities. We hope that this survey will contribute to the ongoing advancement of MLLM research by offering a comprehensive overview of benchmarking practices and identifying promising directions for future work. An associated GitHub repository collecting the latest papers is available.
相關內容
The development of Large Language Models (LLMs) in various languages has been advancing, but the combination of non-English languages with domain-specific contexts remains underexplored. This paper presents our findings from training and evaluating a Japanese business domain-specific LLM designed to better understand business-related documents, such as the news on current affairs, technical reports, and patents. Additionally, LLMs in this domain require regular updates to incorporate the most recent knowledge. Therefore, we also report our findings from the first experiments and evaluations involving updates to this LLM using the latest article data, which is an important problem setting that has not been addressed in previous research. From our experiments on a newly created benchmark dataset for question answering in the target domain, we found that (1) our pretrained model improves QA accuracy without losing general knowledge, and (2) a proper mixture of the latest and older texts in the training data for the update is necessary. Our pretrained model and business domain benchmark are publicly available to support further studies.
The wide deployment of Large Language Models (LLMs) has given rise to strong demands for optimizing their inference performance. Today's techniques serving this purpose primarily focus on reducing latency and improving throughput through algorithmic and hardware enhancements, while largely overlooking their privacy side effects, particularly in a multi-user environment. In our research, for the first time, we discovered a set of new timing side channels in LLM systems, arising from shared caches and GPU memory allocations, which can be exploited to infer both confidential system prompts and those issued by other users. These vulnerabilities echo security challenges observed in traditional computing systems, highlighting an urgent need to address potential information leakage in LLM serving infrastructures. In this paper, we report novel attack strategies designed to exploit such timing side channels inherent in LLM deployments, specifically targeting the Key-Value (KV) cache and semantic cache widely used to enhance LLM inference performance. Our approach leverages timing measurements and classification models to detect cache hits, allowing an adversary to infer private prompts with high accuracy. We also propose a token-by-token search algorithm to efficiently recover shared prompt prefixes in the caches, showing the feasibility of stealing system prompts and those produced by peer users. Our experimental studies on black-box testing of popular online LLM services demonstrate that such privacy risks are completely realistic, with significant consequences. Our findings underscore the need for robust mitigation to protect LLM systems against such emerging threats.
The rapid development of Multimodal Large Language Models (MLLMs) has expanded their capabilities from image comprehension to video understanding. However, most of these MLLMs focus primarily on offline video comprehension, necessitating extensive processing of all video frames before any queries can be made. This presents a significant gap compared to the human ability to watch, listen, think, and respond to streaming inputs in real time, highlighting the limitations of current MLLMs. In this paper, we introduce StreamingBench, the first comprehensive benchmark designed to evaluate the streaming video understanding capabilities of MLLMs. StreamingBench assesses three core aspects of streaming video understanding: (1) real-time visual understanding, (2) omni-source understanding, and (3) contextual understanding. The benchmark consists of 18 tasks, featuring 900 videos and 4,500 human-curated QA pairs. Each video features five questions presented at different time points to simulate a continuous streaming scenario. We conduct experiments on StreamingBench with 13 open-source and proprietary MLLMs and find that even the most advanced proprietary MLLMs like Gemini 1.5 Pro and GPT-4o perform significantly below human-level streaming video understanding capabilities. We hope our work can facilitate further advancements for MLLMs, empowering them to approach human-level video comprehension and interaction in more realistic scenarios.
Large Language Models (LLMs) have demonstrated unparalleled effectiveness in various NLP tasks, and integrating LLMs with automatic speech recognition (ASR) is becoming a mainstream paradigm. Building upon this momentum, our research delves into an in-depth examination of this paradigm on a large open-source Chinese dataset. Specifically, our research aims to evaluate the impact of various configurations of speech encoders, LLMs, and projector modules in the context of the speech foundation encoder-LLM ASR paradigm. Furthermore, we introduce a three-stage training approach, expressly developed to enhance the model's ability to align auditory and textual information. The implementation of this approach, alongside the strategic integration of ASR components, enabled us to achieve the SOTA performance on the AISHELL-1, Test_Net, and Test_Meeting test sets. Our analysis presents an empirical foundation for future research in LLM-based ASR systems and offers insights into optimizing performance using Chinese datasets. We will publicly release all scripts used for data preparation, training, inference, and scoring, as well as pre-trained models and training logs to promote reproducible research.
Large Language Models (LLMs) have shown potential to enhance software development through automated code generation and refactoring, reducing development time and improving code quality. This study empirically evaluates StarCoder2, an LLM optimized for code generation, in refactoring code across 30 open-source Java projects. We compare StarCoder2's performance against human developers, focusing on (1) code quality improvements, (2) types and effectiveness of refactorings, and (3) enhancements through one-shot and chain-of-thought prompting. Our results indicate that StarCoder2 reduces code smells by 20.1% more than developers, excelling in systematic issues like Long Statement and Magic Number, while developers handle complex, context-dependent issues better. One-shot prompting increases the unit test pass rate by 6.15% and improves code smell reduction by 3.52%. Generating five refactorings per input further increases the pass rate by 28.8%, suggesting that combining one-shot prompting with multiple refactorings optimizes performance. These findings provide insights into StarCoder2's potential and best practices for integrating LLMs into software refactoring, supporting more efficient and effective code improvement in real-world applications.
Customer Relationship Management (CRM) systems are vital for modern enterprises, providing a foundation for managing customer interactions and data. Integrating AI agents into CRM systems can automate routine processes and enhance personalized service. However, deploying and evaluating these agents is challenging due to the lack of realistic benchmarks that reflect the complexity of real-world CRM tasks. To address this issue, we introduce CRMArena, a novel benchmark designed to evaluate AI agents on realistic tasks grounded in professional work environments. Following guidance from CRM experts and industry best practices, we designed CRMArena with nine customer service tasks distributed across three personas: service agent, analyst, and manager. The benchmark includes 16 commonly used industrial objects (e.g., account, order, knowledge article, case) with high interconnectivity, along with latent variables (e.g., complaint habits, policy violations) to simulate realistic data distributions. Experimental results reveal that state-of-the-art LLM agents succeed in less than 40% of the tasks with ReAct prompting, and less than 55% even with function-calling abilities. Our findings highlight the need for enhanced agent capabilities in function-calling and rule-following to be deployed in real-world work environments. CRMArena is an open challenge to the community: systems that can reliably complete tasks showcase direct business value in a popular work environment.
Amidst the robust impetus from artificial intelligence (AI) and big data, edge intelligence (EI) has emerged as a nascent computing paradigm, synthesizing AI with edge computing (EC) to become an exemplary solution for unleashing the full potential of AI services. Nonetheless, challenges in communication costs, resource allocation, privacy, and security continue to constrain its proficiency in supporting services with diverse requirements. In response to these issues, this paper introduces socialized learning (SL) as a promising solution, further propelling the advancement of EI. SL is a learning paradigm predicated on social principles and behaviors, aimed at amplifying the collaborative capacity and collective intelligence of agents within the EI system. SL not only enhances the system's adaptability but also optimizes communication, and networking processes, essential for distributed intelligence across diverse devices and platforms. Therefore, a combination of SL and EI may greatly facilitate the development of collaborative intelligence in the future network. This paper presents the findings of a literature review on the integration of EI and SL, summarizing the latest achievements in existing research on EI and SL. Subsequently, we delve comprehensively into the limitations of EI and how it could benefit from SL. Special emphasis is placed on the communication challenges and networking strategies and other aspects within these systems, underlining the role of optimized network solutions in improving system efficiency. Based on these discussions, we elaborate in detail on three integrated components: socialized architecture, socialized training, and socialized inference, analyzing their strengths and weaknesses. Finally, we identify some possible future applications of combining SL and EI, discuss open problems and suggest some future research.
Explainable Artificial Intelligence (XAI) systems aim to improve users' understanding of AI but rarely consider the inclusivity aspects of XAI. Without inclusive approaches, improving explanations might not work well for everyone. This study investigates leveraging users' diverse problem-solving styles as an inclusive strategy to fix an XAI prototype, with the ultimate goal of improving users' mental models of AI. We ran a between-subject study with 69 participants. Our results show that the inclusivity fixes increased participants' engagement with explanations and produced significantly improved mental models. Analyzing differences in mental model scores further highlighted specific inclusivity fixes that contributed to the significant improvement in the mental model. To our surprise, the inclusivity fixes did not improve participants' prediction performance. However, the fixes did improve inclusivity support for women and promoted equity by reducing the gender gap.
The growing demand for personalized decision-making has led to a surge of interest in estimating the Conditional Average Treatment Effect (CATE). Various types of CATE estimators have been developed with advancements in machine learning and causal inference. However, selecting the desirable CATE estimator through a conventional model validation procedure remains impractical due to the absence of counterfactual outcomes in observational data. Existing approaches for CATE estimator selection, such as plug-in and pseudo-outcome metrics, face two challenges. First, they must determine the metric form and the underlying machine learning models for fitting nuisance parameters (e.g., outcome function, propensity function, and plug-in learner). Second, they lack a specific focus on selecting a robust CATE estimator. To address these challenges, this paper introduces a Distributionally Robust Metric (DRM) for CATE estimator selection. The proposed DRM is nuisance-free, eliminating the need to fit models for nuisance parameters, and it effectively prioritizes the selection of a distributionally robust CATE estimator. The experimental results validate the effectiveness of the DRM method in selecting CATE estimators that are robust to the distribution shift incurred by covariate shift and hidden confounders.
The advent of large language models marks a revolutionary breakthrough in artificial intelligence. With the unprecedented scale of training and model parameters, the capability of large language models has been dramatically improved, leading to human-like performances in understanding, language synthesizing, and common-sense reasoning, etc. Such a major leap-forward in general AI capacity will change the pattern of how personalization is conducted. For one thing, it will reform the way of interaction between humans and personalization systems. Instead of being a passive medium of information filtering, large language models present the foundation for active user engagement. On top of such a new foundation, user requests can be proactively explored, and user's required information can be delivered in a natural and explainable way. For another thing, it will also considerably expand the scope of personalization, making it grow from the sole function of collecting personalized information to the compound function of providing personalized services. By leveraging large language models as general-purpose interface, the personalization systems may compile user requests into plans, calls the functions of external tools to execute the plans, and integrate the tools' outputs to complete the end-to-end personalization tasks. Today, large language models are still being developed, whereas the application in personalization is largely unexplored. Therefore, we consider it to be the right time to review the challenges in personalization and the opportunities to address them with LLMs. In particular, we dedicate this perspective paper to the discussion of the following aspects: the development and challenges for the existing personalization system, the newly emerged capabilities of large language models, and the potential ways of making use of large language models for personalization.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191