Knowledge and Reasoning
For AI to be useful, it needs to understand what’s being asked of it, and to know how to respond. At IBM Research, we’re working on systems to help AI better reason with the tasks it’s presented, such as understanding context and analogies, comprehension, and planning through scenarios.
Our work
AI, you have a lot of explaining to do
ReleaseIBM, MIT and Harvard release “Common Sense AI” dataset at ICML 2021
ReleaseIBM’s new AI outperforms competition in table entry search with question-answering
ReleaseGetting AI to reason: using neuro-symbolic AI for knowledge-based question answering
ResearchDualTKB: A Dual Learning Bridge between Text and Knowledge Base
Research
Projects
Accelerating clinical trials
Developing AI and analytics to understand the drivers of study or clinical trial efficiency.
- Accelerated Discovery
- Natural Language Processing
- Machine Learning
- Knowledge and Reasoning
- Healthcare
- Foundation Models
Imprecise probabilistic logic
Developing a general knowledge representation framework to facilitate effective reasoning over multiple sources of imprecise knowledge.
- Neuro-symbolic AI
- Knowledge and Reasoning
- Causality
Extensions and NLU Applications of Logical Neural Networks
Using Logical Neural Networks to demonstrate the benefit of incorporating knowledge and reasoning into neural network learning
- Knowledge and Reasoning
- Neuro-symbolic AI
- Natural Language Processing
- Machine Learning
Publications
- 2023
- AAAI 2023
- 2022
- arXiv
- 2022
- EMNLP 2022
- 2022
- EMNLP 2022
- 2022
- EMNLP 2022
- 2022
- EMNLP 2022