Augument ≠ Replace.

Uchenna Akujuobi et al.

Explores LLM-driven literature-based discovery for scientific hypothesis generation, surfacing novel connections from research literature.

Represents my current research direction: AI systems that accelerate discovery workflows.

Uchenna Akujuobi, J Yi, ME Chung, TR Besold

Presents the full Gastro-Health research agenda: personalized nutrition, GI-aware food recommendation, and health forecasting through integrated AI.

Uchenna Akujuobi, S Liu, TR Besold

Revisits NER methods for the food domain, improving extraction robustness for ingredients, nutrients, and food-related entities.

Uchenna Akujuobi, P Kumari, J Choi, S Badreddine, K Maruyama et al.

Uses temporal graph link prediction with active curriculum learning to generate and rank scientific hypotheses.

L Cabello, Uchenna Akujuobi

Shows that straightforward improvements to aspect-based sentiment analysis can outperform complex recent methods.

L Cabello, C Martin-Turrero, Uchenna Akujuobi, A Søgaard, C Bobed

Augments LLMs with structured medical knowledge graphs to improve factual accuracy in medical question answering.

JC Young, Uchenna Akujuobi

Uses sentiment analysis of citation contexts to improve literature-based discovery and hypothesis ranking.

L Xiao, P Xu, L Jing, Uchenna Akujuobi, X Zhang

Introduces semantic guidance to improve semi-supervised few-shot multi-label node classification in attributed graphs.

Uchenna Akujuobi, M Spranger, SK Palaniappan, X Zhang

Learns temporal node-pair embeddings on biomedical knowledge graphs to rank and generate plausible research hypotheses.

One of my most-cited papers (32 citations) — a core contribution to temporal hypothesis generation.

Uchenna Akujuobi

PhD thesis on graph representation learning for scholarly knowledge, spanning dataset mining, citation networks, and temporal hypothesis generation.

Uchenna Akujuobi, J Chen, M Elhoseiny, M Spranger, X Zhang

Frames hypothesis generation as positive-unlabeled learning with temporal risk estimation over biomedical knowledge graphs.

Published at NeurIPS 2020 — the highest-profile venue in my publication record.

Uchenna Akujuobi, Qiannan Zhang, Han Yufei, Xiangliang Zhang

Models graph-based semi-supervised classification as a learned walk policy that selectively explores node neighborhoods.

A key graph learning contribution from the KAUST research period, published at WSDM 2020.

Uchenna Akujuobi, Han Yufei, Qiannan Zhang, Xiangliang Zhang

Introduces cooperative label-specific graph walks for multi-label node classification in attributed graphs.

B Altaf, Uchenna Akujuobi, L Yu, X Zhang

Uses a variational graph autoencoder to recommend datasets to researchers based on topic and citation context.

Most-cited of my papers (44 citations) — a strong result on dataset-driven scholarly AI.

Uchenna Akujuobi, Ke Sun, Xiangliang Zhang

Extracts popular datasets used across research areas by learning from document content and citation relationships.

Uchenna Akujuobi, Xiangliang Zhang

A dataset-first scholarly search system for finding relevant datasets and analyzing document-dataset citation relationships.

Bridges academic search, dataset retrieval, and visual analysis — built during my KAUST doctoral period.

Uchenna Akujuobi, Xiangliang Zhang

Demo paper presenting Delve as a web system for dataset retrieval and document analysis.