Zheng Li
Publications
Design and Empirical Study of a Large Language Model-Based Multi-Agent Investment System for Chinese Public REITs
This study addresses the low-volatility Chinese Public Real Estate Investment Trusts (REITs) market, proposing a large language model (LLM)-driven trading framework based on multi-agent collaboration. The system constructs four types of analytical agents-announcement, event, price momentum, and market-each conducting analysis from different dimensions; then the prediction agent integrates these multi-source signals to output directional probability distributions across multiple time horizons, then the decision agent generates discrete position adjustment signals based on the prediction results and risk control constraints, thereby forming a closed loop of analysis-prediction-decision-execution. This study further compares two prediction model pathways: for the prediction agent, directly calling the general-purpose large model DeepSeek-R1 versus using a specialized small model Qwen3-8B fine-tuned via supervised fine-tuning and reinforcement learning alignment. In the backtest from October 2024 to October 2025, both agent-based strategies significantly outperformed the buy-and-hold benchmark in terms of cumulative return, Sharpe ratio, and maximum drawdown. The results indicate that the multi-agent framework can effectively enhance the risk-adjusted return of REITs trading, and the fine-tuned small model performs close to or even better than the general-purpose large model in some scenarios.
TIDAL: Temporally Interleaved Diffusion and Action Loop for High-Frequency VLA Control
Large-scale Vision-Language-Action (VLA) models offer semantic generalization but suffer from high inference latency, limiting them to low-frequency batch-and-execute paradigm. This frequency mismatch creates an execution blind spot, causing failures in dynamic environments where targets move during the open-loop execution window. We propose TIDAL (Temporally Interleaved Diffusion and Action Loop), a hierarchical framework that decouples semantic reasoning from high-frequency actuation. TIDAL operates as a backbone-agnostic module for diffusion-based VLAs, using a dual-frequency architecture to redistribute the computational budget. Specifically, a low-frequency macro-intent loop caches semantic embeddings, while a high-frequency micro-control loop interleaves single-step flow integration with execution. This design enables approximately 9 Hz control updates on edge hardware (vs. approximately 2.4 Hz baselines) without increasing marginal overhead. To handle the resulting latency shift, we introduce a temporally misaligned training strategy where the policy learns predictive compensation using stale semantic intent alongside real-time proprioception. Additionally, we address the insensitivity of static vision encoders to velocity by incorporating a differential motion predictor. TIDAL is architectural, making it orthogonal to system-level optimizations. Experiments show a 2x performance gain over open-loop baselines in dynamic interception tasks. Despite a marginal regression in static success rates, our approach yields a 4x increase in feedback frequency and extends the effective horizon of semantic embeddings beyond the native action chunk size. Under non-paused inference protocols, TIDAL remains robust where standard baselines fail due to latency.
DynaDebate: Breaking Homogeneity in Multi-Agent Debate with Dynamic Path Generation
Recent years have witnessed the rapid development of Large Language Model-based Multi-Agent Systems (MAS), which excel at collaborative decision-making and complex problem-solving. Recently, researchers have further investigated Multi-Agent Debate (MAD) frameworks, which enhance the reasoning and collaboration capabilities of MAS through information exchange and debate among multiple agents. However, existing approaches often rely on unguided initialization, causing agents to adopt identical reasoning paths that lead to the same errors. As a result, effective debate among agents is hindered, and the final outcome frequently degenerates into simple majority voting. To solve the above problem, in this paper, we introduce Dynamic Multi-Agent Debate (DynaDebate), which enhances the effectiveness of multi-agent debate through three key mechanisms: (1) Dynamic Path Generation and Allocation, which employs a dedicated Path Generation Agent to generate diverse and logical solution paths with adaptive redundancy; (2) Process-Centric Debate, which shifts the focus from surface-level outcome voting to rigorous step-by-step logic critique to ensure process correctness; (3) A Trigger-Based Verification Agent, which is activated upon disagreement and uses external tools to objectively resolve deadlocks. Extensive experiments demonstrate that DynaDebate achieves superior performance across various benchmarks, surpassing existing state-of-the-art MAD methods.
Key-Value Pair-Free Continual Learner via Task-Specific Prompt-Prototype
Continual learning aims to enable models to acquire new knowledge while retaining previously learned information. Prompt-based methods have shown remarkable performance in this domain; however, they typically rely on key-value pairing, which can introduce inter-task interference and hinder scalability. To overcome these limitations, we propose a novel approach employing task-specific Prompt-Prototype (ProP), thereby eliminating the need for key-value pairs. In our method, task-specific prompts facilitate more effective feature learning for the current task, while corresponding prototypes capture the representative features of the input. During inference, predictions are generated by binding each task-specific prompt with its associated prototype. Additionally, we introduce regularization constraints during prompt initialization to penalize excessively large values, thereby enhancing stability. Experiments on several widely used datasets demonstrate the effectiveness of the proposed method. In contrast to mainstream prompt-based approaches, our framework removes the dependency on key-value pairs, offering a fresh perspective for future continual learning research.