Bridging the Gap Between Theory and Practice of Approximate Bayesian Inference

Johan Kwisthout from Leiden University discusses the challenge of modeling human cognition with Bayesian computations, which are often computationally intractable. The paper analyzes three candidate notions of 'approximation' suggested in cognitive science literature to transform intractable models into computationally plausible ones. It proposes using parameterized computational complexity analyses to derive tractable model variants.

Use Cases

• Developing computationally plausible cognitive models based on the discussed approximation concepts.
• Analyzing the trade-offs between model fidelity and tractability in Bayesian cognitive science.
• Applying parameterized complexity analysis to create variants of intractable Bayesian models.
• Evaluating different notions of 'approximation' for modeling human decision-making and inference.

Strengths

• Authored by a researcher from Leiden University, providing academic credibility.
• Addresses a core theoretical challenge in computational cognitive science: the tractability gap.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

paperswithcode

Collection Method

Likely contains the text of an academic paper discussing theoretical concepts.