Dongmei Zhang
YOU?
Author Swipe
View article: Illustration Layout Generation for Slide Enhancement with Pixel-based Diffusion Model
Illustration Layout Generation for Slide Enhancement with Pixel-based Diffusion Model Open
View article: Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents
Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents Open
The increasing autonomy of LLM agents in handling sensitive communications, accelerated by Model Context Protocol (MCP) and Agent-to-Agent (A2A) frameworks, creates urgent privacy challenges. While recent work reveals significant gaps betw…
View article: Engineering covalent organic frameworks with defect for high-performance immunosensor
Engineering covalent organic frameworks with defect for high-performance immunosensor Open
View article: LettinGo: Explore User Profile Generation for Recommendation System
LettinGo: Explore User Profile Generation for Recommendation System Open
View article: Large Language Models can Deliver Accurate and Interpretable Time Series Anomaly Detection
Large Language Models can Deliver Accurate and Interpretable Time Series Anomaly Detection Open
View article: Transcriptome analysis-based inference of magnolin metabolic pathways in Magnolia: Enhancing Xinyi production and quality management
Transcriptome analysis-based inference of magnolin metabolic pathways in Magnolia: Enhancing Xinyi production and quality management Open
View article: Auto-Test: Learning Semantic-Domain Constraints for Unsupervised Error Detection in Tables
Auto-Test: Learning Semantic-Domain Constraints for Unsupervised Error Detection in Tables Open
Data cleaning is a long-standing challenge in data management. While powerful logic and statistical algorithms have been developed to detect and repair data errors in tables, existing algorithms predominantly rely on domain-experts to firs…
View article: Text2Grad: Reinforcement Learning from Natural Language Feedback
Text2Grad: Reinforcement Learning from Natural Language Feedback Open
Traditional RLHF optimizes language models with coarse, scalar rewards that mask the fine-grained reasons behind success or failure, leading to slow and opaque learning. Recent work augments RL with textual critiques through prompting or r…
View article: Ploutos: Towards Explainable Stock Movement Prediction with Financial Large Language Model
Ploutos: Towards Explainable Stock Movement Prediction with Financial Large Language Model Open
View article: UFO2: The Desktop AgentOS
UFO2: The Desktop AgentOS Open
Recent Computer-Using Agents (CUAs), powered by multimodal large language models (LLMs), offer a promising direction for automating complex desktop workflows through natural language. However, most existing CUAs remain conceptual prototype…
View article: Auto-Test: Learning Semantic-Domain Constraints for Unsupervised Error Detection in Tables
Auto-Test: Learning Semantic-Domain Constraints for Unsupervised Error Detection in Tables Open
Data cleaning is a long-standing challenge in data management. While powerful logic and statistical algorithms have been developed to detect and repair data errors in tables, existing algorithms predominantly rely on domain-experts to firs…
View article: TableLoRA: Low-rank Adaptation on Table Structure Understanding for Large Language Models
TableLoRA: Low-rank Adaptation on Table Structure Understanding for Large Language Models Open
Tabular data are crucial in many fields and their understanding by large language models (LLMs) under high parameter efficiency paradigm is important. However, directly applying parameter-efficient fine-tuning (PEFT) techniques to tabular …
View article: MuDAF: Long-Context Multi-Document Attention Focusing through Contrastive Learning on Attention Heads
MuDAF: Long-Context Multi-Document Attention Focusing through Contrastive Learning on Attention Heads Open
Large Language Models (LLMs) frequently show distracted attention due to irrelevant information in the input, which severely impairs their long-context capabilities. Inspired by recent studies on the effectiveness of retrieval heads in lon…
View article: Learning Identifiable Structures Helps Avoid Bias in DNN-based Supervised Causal Learning
Learning Identifiable Structures Helps Avoid Bias in DNN-based Supervised Causal Learning Open
Causal discovery is a structured prediction task that aims to predict causal relations among variables based on their data samples. Supervised Causal Learning (SCL) is an emerging paradigm in this field. Existing Deep Neural Network (DNN)-…
View article: Enabling Autonomic Microservice Management through Self-Learning Agents
Enabling Autonomic Microservice Management through Self-Learning Agents Open
The increasing complexity of modern software systems necessitates robust autonomic self-management capabilities. While Large Language Models (LLMs) demonstrate potential in this domain, they often face challenges in adapting their general …
View article: Token-level Proximal Policy Optimization for Query Generation
Token-level Proximal Policy Optimization for Query Generation Open
View article: From Reasoning to Answer: Empirical, Attention-Based and Mechanistic Insights into Distilled DeepSeek R1 Models
From Reasoning to Answer: Empirical, Attention-Based and Mechanistic Insights into Distilled DeepSeek R1 Models Open
View article: UFO: A UI-Focused Agent for Windows OS Interaction
UFO: A UI-Focused Agent for Windows OS Interaction Open
View article: Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents
Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents Open
View article: Table-LLM-Specialist: Language Model Specialists for Tables using Iterative Fine-tuning
Table-LLM-Specialist: Language Model Specialists for Tables using Iterative Fine-tuning Open
View article: Study on the Deterioration Mechanisms of Basalt with Fractures Under Freeze-Thaw Cycling
Study on the Deterioration Mechanisms of Basalt with Fractures Under Freeze-Thaw Cycling Open
View article: SECRET: Towards Scalable and Efficient Code Retrieval via Segmented Deep\n Hashing
SECRET: Towards Scalable and Efficient Code Retrieval via Segmented Deep\n Hashing Open
Code retrieval, which retrieves code snippets based on users' natural\nlanguage descriptions, is widely used by developers and plays a pivotal role in\nreal-world software development. The advent of deep learning has shifted the\nretrieval…
View article: Large Language Model-Brained GUI Agents: A Survey
Large Language Model-Brained GUI Agents: A Survey Open
GUIs have long been central to human-computer interaction, providing an intuitive and visually-driven way to access and interact with digital systems. The advent of LLMs, particularly multimodal models, has ushered in a new era of GUI auto…
View article: Token-level Proximal Policy Optimization for Query Generation
Token-level Proximal Policy Optimization for Query Generation Open
Query generation is a critical task for web search engines (e.g. Google, Bing) and recommendation systems. Recently, state-of-the-art query generation methods leverage Large Language Models (LLMs) for their strong capabilities in context u…
View article: Deoxys: A Causal Inference Engine for Unhealthy Node Mitigation in Large-scale Cloud Infrastructure
Deoxys: A Causal Inference Engine for Unhealthy Node Mitigation in Large-scale Cloud Infrastructure Open
The presence of unhealthy nodes in cloud infrastructure signals the potential failure of machines, which can significantly impact the availability and reliability of cloud services, resulting in negative customer experiences. Effectively a…
View article: Nissist: An Incident Mitigation Copilot based on Troubleshooting Guides
Nissist: An Incident Mitigation Copilot based on Troubleshooting Guides Open
Effective incident management is pivotal for the smooth operation of Microsoft cloud services. In order to expedite incident mitigation, service teams gather troubleshooting knowledge into Troubleshooting Guides (TSGs) accessible to On-Cal…
View article: Table-LLM-Specialist: Language Model Specialists for Tables using Iterative Generator-Validator Fine-tuning
Table-LLM-Specialist: Language Model Specialists for Tables using Iterative Generator-Validator Fine-tuning Open
In this work, we propose Table-LLM-Specialist, or Table-Specialist for short, as a new self-trained fine-tuning paradigm specifically designed for table tasks. Our insight is that for each table task, there often exist two dual versions of…
View article: Synthesis of graphene-MnO<sub>2</sub>-poly (2,2’-dithiodianiline) composite for high performance electrochemical electrode materials
Synthesis of graphene-MnO<sub>2</sub>-poly (2,2’-dithiodianiline) composite for high performance electrochemical electrode materials Open
In this study, a hybrid graphene/manganese dioxide/poly(2,2’-dithiodianiline) (graphene/MnO 2 /PDTDA) composite material was in-situ prepared by chemical oxidative polymerization of PDTDA in graphene/MnO 2 and applied as supercapacitor ele…
View article: Pre-trained KPI Anomaly Detection Model Through Disentangled Transformer
Pre-trained KPI Anomaly Detection Model Through Disentangled Transformer Open
In large-scale online service systems, numerous Key Performance Indicators (KPIs), such as service response time and error rate, are gathered in a time-series format. KPI Anomaly Detection (KAD) is a critical data mining problem due to its…
View article: EfficientRAG: Efficient Retriever for Multi-Hop Question Answering
EfficientRAG: Efficient Retriever for Multi-Hop Question Answering Open
Retrieval-augmented generation (RAG) methods encounter difficulties when addressing complex questions like multi-hop queries. While iterative retrieval methods improve performance by gathering additional information, current approaches oft…