Yangqiu Song
YOU?
Author Swipe
View article: Frontiers in Graph Machine Learning for the Large Model Era
Frontiers in Graph Machine Learning for the Large Model Era Open
View article: Examining the Causal Pathways: Digital Marketing Modalities and Their Effects on Contemporary Consumer Behavior
Examining the Causal Pathways: Digital Marketing Modalities and Their Effects on Contemporary Consumer Behavior Open
Technological advancements in digital marketing technologies are transforming how consumers engage-moving away from traditional marketing approaches to data-driven, personalization strategies. The present study aims to examine the influenc…
View article: WebDevJudge: Evaluating (M)LLMs as Critiques for Web Development Quality
WebDevJudge: Evaluating (M)LLMs as Critiques for Web Development Quality Open
The paradigm of LLM-as-a-judge is emerging as a scalable and efficient alternative to human evaluation, demonstrating strong performance on well-defined tasks. However, its reliability in open-ended tasks with dynamic environments and comp…
View article: AtlasKV: Augmenting LLMs with Billion-Scale Knowledge Graphs in 20GB VRAM
AtlasKV: Augmenting LLMs with Billion-Scale Knowledge Graphs in 20GB VRAM Open
Retrieval-augmented generation (RAG) has shown some success in augmenting large language models (LLMs) with external knowledge. However, as a non-parametric knowledge integration paradigm for LLMs, RAG methods heavily rely on external retr…
View article: AutoGraph-R1: End-to-End Reinforcement Learning for Knowledge Graph Construction
AutoGraph-R1: End-to-End Reinforcement Learning for Knowledge Graph Construction Open
Building effective knowledge graphs (KGs) for Retrieval-Augmented Generation (RAG) is pivotal for advancing question answering (QA) systems. However, its effectiveness is hindered by a fundamental disconnect: the knowledge graph (KG) const…
View article: The Cognitive Bandwidth Bottleneck: Shifting Long-Horizon Agent from Planning with Actions to Planning with Schemas
The Cognitive Bandwidth Bottleneck: Shifting Long-Horizon Agent from Planning with Actions to Planning with Schemas Open
Enabling LLMs to effectively operate long-horizon task which requires long-term planning and multiple interactions is essential for open-world autonomy. Conventional methods adopt planning with actions where a executable action list would …
View article: NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents
NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents Open
Large language models are emerging as powerful tools for scientific law discovery, a foundational challenge in AI-driven science. However, existing benchmarks for this task suffer from a fundamental methodological trilemma, forcing a trade…
View article: InteGround: On the Evaluation of Verification and Retrieval Planning in Integrative Grounding
InteGround: On the Evaluation of Verification and Retrieval Planning in Integrative Grounding Open
Grounding large language models (LLMs) in external knowledge sources is a promising method for faithful prediction. While existing grounding approaches work well for simple queries, many real-world information needs require synthesizing mu…
View article: Structuring the Unstructured: A Systematic Review of Text-to-Structure Generation for Agentic AI with a Universal Evaluation Framework
Structuring the Unstructured: A Systematic Review of Text-to-Structure Generation for Agentic AI with a Universal Evaluation Framework Open
The evolution of AI systems toward agentic operation and context-aware retrieval necessitates transforming unstructured text into structured formats like tables, knowledge graphs, and charts. While such conversions enable critical applicat…
View article: Prospect Theory Fails for LLMs: Revealing Instability of Decision-Making under Epistemic Uncertainty
Prospect Theory Fails for LLMs: Revealing Instability of Decision-Making under Epistemic Uncertainty Open
Prospect Theory (PT) models human decision-making under uncertainty, while epistemic markers (e.g., maybe) serve to express uncertainty in language. However, it remains largely unexplored whether Prospect Theory applies to contemporary Lar…
View article: EFO <i> <sub>k</sub> </i> -CQA: Towards Knowledge Graph Complex Query Answering beyond Set Operation
EFO <i> <sub>k</sub> </i> -CQA: Towards Knowledge Graph Complex Query Answering beyond Set Operation Open
View article: From Web Search towards Agentic Deep Research: Incentivizing Search with Reasoning Agents
From Web Search towards Agentic Deep Research: Incentivizing Search with Reasoning Agents Open
Information retrieval is a cornerstone of modern knowledge acquisition, enabling billions of queries each day across diverse domains. However, traditional keyword-based search engines are increasingly inadequate for handling complex, multi…
View article: XToM: Exploring the Multilingual Theory of Mind for Large Language Models
XToM: Exploring the Multilingual Theory of Mind for Large Language Models Open
Theory of Mind (ToM), the ability to infer mental states in others, is pivotal for human social cognition. Existing evaluations of ToM in LLMs are largely limited to English, neglecting the linguistic diversity that shapes human cognition.…
View article: SwitchLingua: The First Large-Scale Multilingual and Multi-Ethnic Code-Switching Dataset
SwitchLingua: The First Large-Scale Multilingual and Multi-Ethnic Code-Switching Dataset Open
Code-switching (CS) is the alternating use of two or more languages within a conversation or utterance, often influenced by social context and speaker identity. This linguistic phenomenon poses challenges for Automatic Speech Recognition (…
View article: Controllable Logical Hypothesis Generation for Abductive Reasoning in Knowledge Graphs
Controllable Logical Hypothesis Generation for Abductive Reasoning in Knowledge Graphs Open
Abductive reasoning in knowledge graphs aims to generate plausible logical hypotheses from observed entities, with broad applications in areas such as clinical diagnosis and scientific discovery. However, due to a lack of controllability, …
View article: INFERENCEDYNAMICS: Efficient Routing Across LLMs through Structured Capability and Knowledge Profiling
INFERENCEDYNAMICS: Efficient Routing Across LLMs through Structured Capability and Knowledge Profiling Open
Large Language Model (LLM) routing is a pivotal technique for navigating a diverse landscape of LLMs, aiming to select the best-performing LLMs tailored to the domains of user queries, while managing computational resources. However, curre…
View article: EcomScriptBench: A Multi-task Benchmark for E-commerce Script Planning via Step-wise Intention-Driven Product Association
EcomScriptBench: A Multi-task Benchmark for E-commerce Script Planning via Step-wise Intention-Driven Product Association Open
Goal-oriented script planning, or the ability to devise coherent sequences of actions toward specific goals, is commonly employed by humans to plan for typical activities. In e-commerce, customers increasingly seek LLM-based assistants to …
View article: Legal Rule Induction: Towards Generalizable Principle Discovery from Analogous Judicial Precedents
Legal Rule Induction: Towards Generalizable Principle Discovery from Analogous Judicial Precedents Open
Legal rules encompass not only codified statutes but also implicit adjudicatory principles derived from precedents that contain discretionary norms, social morality, and policy. While computational legal research has advanced in applying e…
View article: MCIP: Protecting MCP Safety via Model Contextual Integrity Protocol
MCIP: Protecting MCP Safety via Model Contextual Integrity Protocol Open
As Model Context Protocol (MCP) introduces an easy-to-use ecosystem for users and developers, it also brings underexplored safety risks. Its decentralized architecture, which separates clients and servers, poses unique challenges for syste…
View article: Efficient and Scalable Neural Symbolic Search for Knowledge Graph Complex Query Answering
Efficient and Scalable Neural Symbolic Search for Knowledge Graph Complex Query Answering Open
Complex Query Answering (CQA) aims to retrieve answer sets for complex logical formulas from incomplete knowledge graphs, which is a crucial yet challenging task in knowledge graph reasoning. While neuro-symbolic search utilized neural lin…
View article: Transformers for Complex Query Answering over Knowledge Hypergraphs
Transformers for Complex Query Answering over Knowledge Hypergraphs Open
Complex Query Answering (CQA) has been extensively studied in recent years. In order to model data that is closer to real-world distribution, knowledge graphs with different modalities have been introduced. Triple KGs, as the classic KGs c…
View article: Can LLMs Generate Tabular Summaries of Science Papers? Rethinking the Evaluation Protocol
Can LLMs Generate Tabular Summaries of Science Papers? Rethinking the Evaluation Protocol Open
Literature review tables are essential for summarizing and comparing collections of scientific papers. We explore the task of generating tables that best fulfill a user's informational needs given a collection of scientific papers. Buildin…
View article: Simulate and Eliminate: Revoke Backdoors for Generative Large Language Models
Simulate and Eliminate: Revoke Backdoors for Generative Large Language Models Open
With rapid advances, generative large language models (LLMs) dominate various Natural Language Processing (NLP) tasks from understanding to reasoning. Yet, language models' inherent vulnerabilities may be exacerbated due to increased acces…
View article: PrivaCI-Bench: Evaluating Privacy with Contextual Integrity and Legal Compliance
PrivaCI-Bench: Evaluating Privacy with Contextual Integrity and Legal Compliance Open
Recent advancements in generative large language models (LLMs) have enabled wider applicability, accessibility, and flexibility. However, their reliability and trustworthiness are still in doubt, especially for concerns regarding individua…
View article: Patterns Over Principles: The Fragility of Inductive Reasoning in LLMs under Noisy Observations
Patterns Over Principles: The Fragility of Inductive Reasoning in LLMs under Noisy Observations Open
Inductive reasoning, a cornerstone of human cognition, enables generalization from limited data but hasn't yet been fully achieved by large language models (LLMs). While modern LLMs excel at reasoning tasks, their ability to maintain stabl…
View article: Top Ten Challenges Towards Agentic Neural Graph Databases
Top Ten Challenges Towards Agentic Neural Graph Databases Open
Graph databases (GDBs) like Neo4j and TigerGraph excel at handling interconnected data but lack advanced inference capabilities. Neural Graph Databases (NGDBs) address this by integrating Graph Neural Networks (GNNs) for predictive analysi…
View article: LogiDynamics: Unraveling the Dynamics of Inductive, Abductive and Deductive Logical Inferences in LLM Reasoning
LogiDynamics: Unraveling the Dynamics of Inductive, Abductive and Deductive Logical Inferences in LLM Reasoning Open
Modern large language models (LLMs) employ diverse logical inference mechanisms for reasoning, making the strategic optimization of these approaches critical for advancing their capabilities. This paper systematically investigate the compa…
View article: Bridging the Gap Between LLMs and Human Intentions: Progresses and Challenges in Instruction Understanding, Intention Reasoning, and Reliable Generation
Bridging the Gap Between LLMs and Human Intentions: Progresses and Challenges in Instruction Understanding, Intention Reasoning, and Reliable Generation Open
Large language models (LLMs) have demonstrated exceptional capabilities in understanding and generation. However, when interacting with human instructions in real-world scenarios, LLMs still face significant challenges, particularly in acc…
View article: Top Ten Challenges Towards Agentic Neural Graph Databases
Top Ten Challenges Towards Agentic Neural Graph Databases Open
Graph databases (GDBs) like Neo4j and TigerGraph excel at handling interconnected data but lack advanced inference capabilities. Neural Graph Databases (NGDBs) address this by integrating Graph Neural Networks (GNNs) for predictive analysi…
View article: DivScene: Towards Open-Vocabulary Object Navigation with Large Vision Language Models in Diverse Scenes
DivScene: Towards Open-Vocabulary Object Navigation with Large Vision Language Models in Diverse Scenes Open