Human Inductive Bias

Human inductive bias refers to the inherent assumptions and preferences that shape how humans learn from limited data, enabling rapid generalization and efficient knowledge acquisition. Current research focuses on computationally modeling these biases, often employing contrastive learning, Bayesian methods, and program synthesis techniques to capture human-like learning strategies in machine learning models, particularly within the context of few-shot learning and efficient representation of complex systems like language. Understanding and replicating these biases is crucial for improving the robustness, generalizability, and alignment of artificial intelligence systems, bridging the gap between human and machine learning capabilities and addressing reproducibility concerns in both fields.

Papers

May 29, 2024

Using Contrastive Learning with Generative Similarity to Learn Spaces that Capture Human Inductive Biases
Raja Marjieh, Sreejan Kumar, Declan Campbell, Liyi Zhang, Gianluca Bencomo, Jake Snell, Thomas L. Griffiths
Contrastive Learning Inductive Bias Generative Learning Mixed Space Complex Generative Human Inductive Bias

May 11, 2024

Finding structure in logographic writing with library learning
Guangyuan Jiang, Matthias Hofer, Jiayuan Mao, Lionel Wong, Joshua B. Tenenbaum, Roger P. Levy
Human Language Efficient Representation Linguistic Structure Structure Discovery Combinatorial Structure Logographic Writing System Human Inductive Bias

May 23, 2022

Using Natural Language and Program Abstractions to Instill Human Inductive Biases in Machines
Sreejan Kumar, Carlos G. Correa, Ishita Dasgupta, Raja Marjieh, Michael Y. Hu, Robert D. Hawkins, Nathaniel D. Daw, Jonathan D. Cohen, Karthik Narasimhan, Thomas L. Griffiths
Natural Language New Machine Program Induction Human Inductive Bias Programming Abstraction

March 12, 2022

The worst of both worlds: A comparative analysis of errors in learning from data in psychology and machine learning
Jessica Hullman, Sayash Kapoor, Priyanka Nanayakkara, Andrew Gelman, Arvind Narayanan
Machine Learning Raw Data Comparative Study Error Feedback Psychological Phenomenon Causal Attribution Human Inductive Bias

Human Inductive Bias

Papers

Using Contrastive Learning with Generative Similarity to Learn Spaces that Capture Human Inductive Biases

Finding structure in logographic writing with library learning

Using Natural Language and Program Abstractions to Instill Human Inductive Biases in Machines

The worst of both worlds: A comparative analysis of errors in learning from data in psychology and machine learning