Summary of What Makes Models Compositional? a Theoretical View: with Supplement, by Parikshit Ram and Tim Klinger and Alexander G. Gray
What makes Models Compositional? A Theoretical View: With Supplement
by Parikshit Ram, Tim Klinger, Alexander G. Gray
First submitted to arxiv on: 2 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel paper investigates the role of compositional structure in sequence processing models’ failures to generalize on compositional benchmarks. The authors propose a neuro-symbolic definition of compositional functions and analyze various models’ expressivity, sample complexity, and compositional complexity. They provide theoretical guarantees for compositional models that explicitly depend on this definition, shedding light on factors driving poor empirical performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper explores why existing sequence processing models struggle with compositional generalization. The researchers define a new way to understand compositional functions and analyze how different types of models (like recurrent and attention-based ones) use this structure. They show that some models are better than others at handling complex sequences, and they identify factors that make it hard for models to generalize well. |
Keywords
» Artificial intelligence » Attention » Generalization
