8 papers
RAGStress: A controlled benchmark for evaluating retrieval-augmented generation under knowledge-base degradation
Shiqi Yang, Jiekai Ma, Gaoyuan Du
Retrieval-Augmented Generation (RAG) is typically evaluated under the implicit assumption that the underlying knowledge base (KB) is clean, leaving the behaviour of RAG systems und…
Greedy Decoding Is Not Precision-Invariant: Cross-Precision Output Divergence in LLM Inference
Gaoyuan Du, Anam Nawaz Khan, Rex Zhou +4
Greedy decoding from large language models is commonly treated as deterministic. We show it is not precision-invariant: the same model, prompt, and decoding algorithm produce diffe…
When Does Context Routing Help? A Systematic Study of Multi-Modal Fusion in Time Series Forecasting
Ruizhe Zhou, Gaoyuan Du, Xiaoyang Liu +4
Multi-modal time series forecasting methods integrate auxiliary context into temporal predictions through increasingly sophisticated fusion mechanisms. A growing body of work repor…
Carbon-Aware Compute--Power Scheduling for AI Data Centers with Microgrid Prosumer Operations
Johnny R. Zhang, Gaoyuan Du, Qianyi Sun +3
AI data centers are increasingly becoming tightly coupled compute--energy systems, where workload placement, cooling demand, electricity procurement, storage operation, and carbon…
From Accuracy to Auditability: A Survey of Determinism in Financial AI Systems
Ruizhe Zhou, Xiaoyang Liu, Gaoyuan Du +4
Deploying machine learning in regulated financial environments -- credit risk, fraud detection, and anti-money laundering -- exposes critical vulnerabilities in algorithmic reprodu…
A Framework for Assessing AI Agent Decisions and Outcomes in AutoML Pipelines
Gaoyuan Du, Amit Ahlawat, Xiaoyang Liu +1
Agent-based AutoML systems rely on large language models to make complex, multi-stage decisions across data processing, model selection, and evaluation. However, existing evaluatio…
Cross-Modal Memory Compression for Efficient Multi-Agent Debate
Jing Wu, Yue Sun, Tianpei Xie +7
Multi-agent debate can improve reasoning quality and reduce hallucinations, but it incurs rapidly growing context as debate rounds and agent count increase. Retaining full textual…
A Universal Banach--Bregman Framework for Stochastic Iterations: Unifying Stochastic Mirror Descent, Learning and LLM Training
Johnny R. Zhang, Xiaomei Mi, Gaoyuan Du +4
Stochastic optimization powers the scalability of modern artificial intelligence, spanning machine learning, deep learning, reinforcement learning, and large language model trainin…