Cambridge Healthtech Institute’s 2nd Annual

Data Strategies and the Future of AI Models

Empowering Machine Learning with Smarter Data and Next-Generation Tools

January 20-22, 2026

For 2026, the three-day Data Strategies and the Future of AI Models track will explore the essential relationship between wet lab experiments and data science, and then look forward to new models and capabilities that will extend the value of AI and ML in biopharmaceutical R&D. Presenters will offer best practices for the development and acquisition of training data to ensure the best possible results in executing models and consider the range of responses to situations where available training data is insufficient to power experiments for challenging projects. For models, speakers will offer use cases for near term innovations and improvements projected to have an impact in the coming year, then look at longer term goals for models that will extend the use of ML and AI to a wider range of modalities and patient impacts.

Coverage will include, but is not limited to:

Training Data Acquisition and Generation

“ML ready” considerations in assay development and experimental design
Benchmarking public models and datasets
Best practices for the interface of wet lab and data science
Essential assays for training data generation
Iterative data generation with active learning/lab-in-the-loop approaches
Leveraging negative data to improve model specificity and reduce false positives
Technologies for high throughput data generation and validation

Strategies to Overcome Data Limitations

Application of synthetic data and structures
Federated data as a solution for data gaps; barriers and solutions
Generation and sourcing of data for optimization and design of complex biologics
The role and limitations of diffusion models
Working with sparse data for essential properties (PK, bioavailability, immunogenicity)

Near Term Improvements in Models and Capabilities

Co-scientist/agentic models
Emerging model architectures (graph neural networks, transformers, causal inference)
Multispecific and conjugate optimization
Structure-based antibody-antigen complex prediction

Longer Term Wish List

AI based physics models and MD simulations
Modeling system level responses in cells, tissue, and organisms
Models of disease and treatment
Multi-modal/parameter models incorporating sequence, structure and epitope information
Multispecific and conjugate design

Problems and Solutions (presentations or roundtable discussion themes)

Collection and generation of reliable “human” data
Consortium efforts to develop shared datasets
How has ML/AI impacted speed and quality of discovery and engineering workflows?
Pretraining datasets and models
Reducing active learning cycles and cycle times
Strategies for using legacy data
The future course of miniproteins

We are also seeking facilitators to lead interactive elements in addition to the podium presentations.

The deadline for priority consideration is July 11, 2025.

All proposals are subject to review by session chairpersons and/or the Scientific Advisory Committee to ensure the overall quality of the conference program. Additionally, as per Cambridge Healthtech Institute’s policy, a select number of vendors and consultants who provide products and services will be offered opportunities for podium presentation slots based on a variety of Corporate Sponsorships.

Opportunities for Participation: