Showing 1โ€“8 of 12,847 papers
Preprint

Contrastive Fact Verification: Leveraging Adversarial Examples for Robust Claim Validation

Marcus Chen, Dr. Sarah Whitfield โ€” Aevum Preprint Series

We propose a contrastive learning approach to automated fact verification that uses adversarially generated counter-examples to improve model robustness. Our method trains verifiers to distinguish between subtle paraphrases of true and false claims, achieving a 23% improvement on cross-domain generalization benchmarks.

๐Ÿ“… Nov 2024 ๐Ÿ“Š 87 citations ๐Ÿ‘ 5.2K views
Fact Checking NLP Adversarial ML
Conference

Epistemic Equity: Measuring and Mitigating Geographic Bias in Crowdsourced Knowledge Systems

Dr. Fatima Al-Rashid, Prof. James Okonkwo, Dr. Priya Sharma โ€” ACL 2024

This study presents a comprehensive audit of geographic representation across major crowdsourced knowledge platforms. We develop an epistemic equity index and demonstrate systematic underrepresentation of content from Global South perspectives, proposing algorithmic and community interventions to address these gaps.

๐Ÿ“… Oct 2024 ๐Ÿ”— DOI: 10.18653/v1.2024.acl-1.042 ๐Ÿ“Š 156 citations ๐Ÿ‘ 9.8K views
Bias Equity Crowdsourcing
Journal

Temporal Knowledge Representation: Modeling the Evolution of Scientific Concepts Across Decades

Prof. Wei Zhang, Dr. Lisa Andersson โ€” Aevum Research Journal, Vol. 13

We introduce a temporal knowledge representation framework that models how scientific concepts evolve over time. Using a corpus of 2 million encyclopedia articles spanning 50 years, we trace the development of 14,000+ concepts and identify patterns of conceptual drift, convergence, and paradigm shifts.

๐Ÿ“… Sep 2024 ๐Ÿ“Š 93 citations ๐Ÿ‘ 4.1K views
Temporal KG Scientometrics
Preprint

Low-Resource Language Support in Semantic Search: A Zero-Shot Transfer Approach Using Cross-Lingual Embeddings

Dr. Kwame Asante, Nina Petrov โ€” Aevum Preprint Series

Addressing the language gap in semantic search, we present a zero-shot transfer framework using cross-lingual embeddings that supports 87 low-resource languages without any language-specific training data. Our approach leverages massively multilingual models combined with adaptive retrieval to achieve competitive search quality.

๐Ÿ“… Aug 2024 ๐Ÿ“Š 64 citations ๐Ÿ‘ 3.7K views
Low-Resource Semantic Search Cross-Lingual
Thesis

Causal Reasoning in Natural Language Explanations: A Framework for Automated Explanation Quality Assessment

Alex Rivera โ€” PhD Thesis, Stanford University, 2024

This thesis develops a causal reasoning framework for evaluating the quality of natural language explanations in educational and encyclopedia contexts. We introduce the Explanatory Causal Fidelity (ECF) metric and demonstrate its correlation with human judgments of explanation quality across diverse domains.

๐Ÿ“… Jul 2024 ๐Ÿ“Š 41 citations ๐Ÿ‘ 2.3K views
Causal Reasoning Explanations
Conference

Federated Knowledge Editing: Collaborative Update Mechanisms for Distributed Encyclopedia Systems

Dr. Yuki Sato, Prof. David Mueller, Dr. Ingrid Larsen โ€” WWW 2024

We propose a federated editing framework that enables consistent, conflict-free knowledge updates across distributed encyclopedia instances. Our protocol achieves 99.7% consistency while supporting concurrent edits from thousands of contributors, with real-time conflict resolution using version vector lattices.

๐Ÿ“… Jun 2024 ๐Ÿ”— DOI: 10.1145/3589334.3645612 ๐Ÿ“Š 118 citations ๐Ÿ‘ 6.5K views
Federated Distributed Systems
Journal

Neurosymbolic Knowledge Acquisition: Combining Neural Language Models with Formal Ontology Reasoning

Prof. Maria Santos, Dr. Thomas Wright โ€” Aevum Research Journal, Vol. 12

This paper bridges the gap between neural language understanding and symbolic reasoning for knowledge acquisition. We present a neurosymbolic architecture that uses LLMs for information extraction and description logic reasoners for consistency validation, achieving both scalability and formal guarantees.

๐Ÿ“… May 2024 ๐Ÿ“Š 76 citations ๐Ÿ‘ 4.8K views
Neurosymbolic Ontology Reasoning