Jinyu Cai

Jinyu Cai, Jialong Li, Nianyu Li et al.

The rapid iteration cycles of modern live-service games make regression testing indispensable for maintaining quality and stability. However, existing regression testing approaches face critical limitations, especially in common gray-box settings where full source code access is unavailable: they heavily rely on manual effort for test case construction, struggle to maintain growing suites plagued by redundancy, and lack efficient mechanisms for prioritizing relevant tests. These challenges result in excessive testing costs, limited automation, and insufficient bug detection. To address these issues, we propose SAGE, a semanticaware regression testing framework for gray-box game environments. SAGE systematically addresses the core challenges of test generation, maintenance, and selection. It employs LLM-guided reinforcement learning for efficient, goal-oriented exploration to automatically generate a diverse foundational test suite. Subsequently, it applies a semantic-based multi-objective optimization to refine this suite into a compact, high-value subset by balancing cost, coverage, and rarity. Finally, it leverages LLM-based semantic analysis of update logs to prioritize test cases most relevant to version changes, enabling efficient adaptation across iterations. We evaluate SAGE on two representative environments, Overcooked Plus and Minecraft, comparing against both automated baselines and human-recorded test cases. Across all environments, SAGE achieves superior bug detection with significantly lower execution cost, while demonstrating strong adaptability to version updates.

Jinyu Cai, Jialong Li, Mingyue Zhang et al.

Social media platforms such as Twitter, Reddit, and Sina Weibo play a crucial role in global communication but often encounter strict regulations in geopolitically sensitive regions. This situation has prompted users to ingeniously modify their way of communicating, frequently resorting to coded language in these regulated social media environments. This shift in communication is not merely a strategy to counteract regulation, but a vivid manifestation of language evolution, demonstrating how language naturally evolves under societal and technological pressures. Studying the evolution of language in regulated social media contexts is of significant importance for ensuring freedom of speech, optimizing content moderation, and advancing linguistic research. This paper proposes a multi-agent simulation framework using Large Language Models (LLMs) to explore the evolution of user language in regulated social media environments. The framework employs LLM-driven agents: supervisory agent who enforce dialogue supervision and participant agents who evolve their language strategies while engaging in conversation, simulating the evolution of communication styles under strict regulations aimed at evading social media regulation. The study evaluates the framework's effectiveness through a range of scenarios from abstract scenarios to real-world situations. Key findings indicate that LLMs are capable of simulating nuanced language dynamics and interactions in constrained settings, showing improvement in both evading supervision and information accuracy as evolution progresses. Furthermore, it was found that LLM agents adopt different strategies for different scenarios.

Zichen Zhu, Yang Xu, Lu Chen et al.

The rapid development of multimodal large language models (MLLMs) raises the question of how they compare to human performance. While existing datasets often feature synthetic or overly simplistic tasks, some models have already surpassed human expert baselines. In this paper, we present MULTI, a Chinese multimodal dataset derived from authentic examination questions. Comprising over 18,000 carefully selected and refined questions, MULTI evaluates models using real-world examination standards, encompassing image-text comprehension, complex reasoning, and knowledge recall. Additionally, We also introduce MULTI-Elite, a 500-question selected hard subset, and MULTI-Extend with more than 4,500 external knowledge context pieces for testing in-context learning capabilities. Our evaluation highlights substantial room for MLLM advancement, with Qwen2-VL-72B achieving a 76.9% accuracy on MULTI and 53.1% on MULTI-Elite leading 25 evaluated models, compared to human expert baselines of 86.1% and 73.1%. MULTI serves not only as a robust evaluation platform but also paves the way for the development of expert-level AI.

Jinyu Cai, Yusei Ishimizu, Mingyue Zhang et al.

Social media platforms frequently impose restrictive policies to moderate user content, prompting the emergence of creative evasion language strategies. This paper presents a multi-agent framework based on Large Language Models (LLMs) to simulate the iterative evolution of language strategies under regulatory constraints. In this framework, participant agents, as social media users, continuously evolve their language expression, while supervisory agents emulate platform-level regulation by assessing policy violations. To achieve a more faithful simulation, we employ a dual design of language strategies (constraint and expression) to differentiate conflicting goals and utilize an LLM-driven GA (Genetic Algorithm) for the selection, mutation, and crossover of language strategies. The framework is evaluated using two distinct scenarios: an abstract password game and a realistic simulated illegal pet trade scenario. Experimental results demonstrate that as the number of dialogue rounds increases, both the number of uninterrupted dialogue turns and the accuracy of information transmission improve significantly. Furthermore, a user study with 40 participants validates the real-world relevance of the generated dialogues and strategies. Moreover, ablation studies validate the importance of the GA, emphasizing its contribution to long-term adaptability and improved overall results.

Wei Guan, Jian Cao, Jinyu Cai et al.

Agentic Workflows (AWs) have emerged as a promising paradigm for solving complex tasks. However, the scalability of automating their generation is severely constrained by the high cost and latency of execution-based evaluation. Existing AW performance prediction methods act as surrogates but fail to simultaneously capture the intricate topological dependencies and the deep semantic logic embedded in AWs. To address this limitation, we propose GLOW, a unified framework for AW performance prediction that combines the graph-structure modeling capabilities of GNNs with the reasoning power of LLMs. Specifically, we introduce a graph-oriented LLM, instruction-tuned on graph tasks, to extract topologically aware semantic features, which are fused with GNN-encoded structural representations. A contrastive alignment strategy further refines the latent space to distinguish high-quality AWs. Extensive experiments on FLORA-Bench show that GLOW outperforms state-of-the-art baselines in prediction accuracy and ranking utility.

Jinyu Cai, Jialong Li, Mingyue Zhang et al.

In this paper, we address the challenges faced by Value Iteration Networks (VIN) in handling larger input maps and mitigating the impact of accumulated errors caused by increased iterations. We propose a novel approach, Value Iteration Networks with Gated Summarization Module (GS-VIN), which incorporates two main improvements: (1) employing an Adaptive Iteration Strategy in the Value Iteration module to reduce the number of iterations, and (2) introducing a Gated Summarization module to summarize the iterative process. The adaptive iteration strategy uses larger convolution kernels with fewer iteration times, reducing network depth and increasing training stability while maintaining the accuracy of the planning process. The gated summarization module enables the network to emphasize the entire planning process, rather than solely relying on the final global planning outcome, by temporally and spatially resampling the entire planning process within the VI module. We conduct experiments on 2D grid world path-finding problems and the Atari Mr. Pac-man environment, demonstrating that GS-VIN outperforms the baseline in terms of single-step accuracy, planning success rate, and overall performance across different map sizes. Additionally, we provide an analysis of the relationship between input size, kernel size, and the number of iterations in VI-based models, which is applicable to a majority of VI-based models and offers valuable insights for researchers and industrial deployment.