Summary of The Kandy Benchmark: Incremental Neuro-symbolic Learning and Reasoning with Kandinsky Patterns, by Luca Salvatore Lorello et al.
The KANDY Benchmark: Incremental Neuro-Symbolic Learning and Reasoning with Kandinsky Patterns
by Luca Salvatore Lorello, Marco Lippi, Stefano Melacci
First submitted to arxiv on: 27 Feb 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Machine Learning (cs.LG)
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 The proposed KANDY benchmarking framework generates diverse learning and reasoning tasks inspired by Kandinsky patterns, creating curricula of binary classification tasks with increasing complexity and sparse supervisions. This enables the implementation of benchmarks for continual and semi-supervised learning, focusing on symbol compositionality. The framework provides classification rules in the ground truth for interpretable solution analysis. Two curricula are released: an easier and a harder one, serving as new challenges for the research community. Experimental evaluation reveals that both neural models and symbolic approaches struggle to solve most tasks, highlighting the need for advanced neuro-symbolic methods trained over time. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary KANDY is a new way to test AI systems. It creates many different problems for machines to learn from, using patterns inspired by Kandinsky’s art. This helps researchers see how well different AI approaches can handle learning and reasoning tasks with increasing difficulty. The framework also provides answers in the ground truth so that people can analyze why certain solutions work or don’t. Two sets of challenges are offered: easy and hard. While state-of-the-art AI models and simple symbolic methods struggle to solve most problems, this highlights the need for new approaches that combine neural networks and symbols to improve over time. |
Keywords
* Artificial intelligence * Classification * Semi supervised
