Research

Hill Research Presents at AAAI 2026 SPARTA Workshop in Singapore

Hill Research |
Hill Research Presents at AAAI 2026 SPARTA Workshop in Singapore

Hill Research’s AI team presented three research projects at the SPARTA workshop during the 40th AAAI Conference on Artificial Intelligence (AAAI-26) in Singapore on January 26, 2026. CEO Dr. Louise Liu also delivered a keynote address as a confirmed speaker.

About AAAI 2026

The AAAI Conference on Artificial Intelligence is one of the premier international academic conferences in AI, bringing together researchers and practitioners from around the world. AAAI-26 was held January 20-27, 2026 at the Singapore EXPO.

The SPARTA Workshop

SPARTA (Spatial Reasoning and Therapeutics with AI: From Omics to Imaging) is a one-day workshop focused on the intersection of AI with spatial biology, digital pathology, drug discovery, and clinical trial support. The workshop brought together multi-modal AI researchers, spatial biology experts, radiologists, pathologists, and pharmaceutical data scientists.

Keynote speakers included researchers from Sorbonne University, Dana-Farber Cancer Institute / Harvard, A*STAR Singapore, and Hill Research.

Hill Research Presentations

CSLAN: Cross-Species Latent Alignment Network

Presented by Dr. Rui Wu, AI Lead at Hill Research, CSLAN is a transfer learning framework that bridges the gap between mouse and human single-cell datasets. The approach achieved over 95% accuracy with only 240 human samples by addressing data scarcity and overfitting challenges in translational biology. This work builds on Hill Research’s earlier preprint published on BioRxiv.

Dynamic Consistency Index (DCI)

Also presented by Dr. Rui Wu, the Dynamic Consistency Index is a new metric for measuring how gene expression evolves across immune cell types over time. DCI enables modeling of gene expression dynamics with high accuracy while estimating uncertainty, supporting more robust biological insights.

Consensus-Based Framework for Reducing LLM Hallucinations

Presented by Zhiyuan Chen, this research addresses the critical challenge of hallucinations in large language models within clinical settings. Rather than relying on single judge models or simple matching techniques, the team developed a consensus-based framework where models from different families challenge each other. The approach improves accuracy while advancing toward more transparent and audit-ready clinical AI systems.

Why It Matters

All three projects reflect Hill Research’s commitment to building trustworthy, audit-ready AI for healthcare and clinical trials. The research spans from foundational biology (cross-species transfer learning, gene expression dynamics) to applied clinical AI (reducing LLM hallucinations), demonstrating the breadth of the team’s capabilities.

As Dr. Rui Wu emphasized, the focus is on “robust knowledge transfer and dynamic modeling for audit-ready, trustworthy AI in biology.”

Learn More