Knowledge representation · Graph-grounded AI · Cultural heritage
Russa Biswas
Tenure-Track Assistant Professor · Co-director, AI:PAGE-Lab
Russa studies how knowledge graphs represent information and support reasoning, including how they can make language-model outputs more factual.
She develops methods for knowledge-graph completion and entity typing, combining graph structure with textual descriptions. Her recent work examines multilingual hallucinations and reasoning supported by explicit graph paths.
Russa co-directs AI:PAGE-Lab with historian Johan Heinsen. The lab turns historical texts into searchable, connected knowledge so that researchers can explore relationships across cultural-heritage collections.
Knowledge representation
Graph embeddings, entity typing, link prediction, and completion methods that combine graph structure with textual descriptions.
Factual and informative language models
Multilingual hallucination evaluation, factual completeness, and reasoning grounded in explicit knowledge-graph paths.
Knowledge across languages and domains
Multilingual and lower-resourced knowledge graphs, and connected historical knowledge through AI:PAGE-Lab.
Selected work
5 representative papers
MultiHal: Multilingual Dataset for Knowledge-Graph Grounded Evaluation of LLM Hallucinations
Builds a multilingual multihop benchmark from 25,900 curated knowledge-graph paths for evaluating and reducing hallucinations.
- Knowledge, Evidence & Reasoning
- Multilingual & Lower-Resource NLP
Knowledge Graphs, Large Language Models, and Hallucinations: An NLP Perspective
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
Multilingual Knowledge Graphs and Low-Resource Languages: A Review
- Knowledge, Evidence & Reasoning
- Multilingual & Lower-Resource NLP
Knowledge Graph Embeddings: Open Challenges and Opportunities
- Knowledge, Evidence & Reasoning
- Factuality, Reliability & Evaluation
MADLINK: Attentive multihop and entity descriptions for link prediction in knowledge graphs
Combines graph paths with entity descriptions to predict missing links in knowledge graphs.
- Knowledge, Evidence & Reasoning
Supervision and collaboration
Research with Russa Biswas
Russa welcomes projects on knowledge-graph representation and reasoning, multilingual factuality, and applications of structured knowledge.