Summary of Coverage-constrained Human-ai Cooperation with Multiple Experts, by Zheng Zhang et al.
Coverage-Constrained Human-AI Cooperation with Multiple Experts
by Zheng Zhang, Cuong Nguyen, Kevin Wells, Thanh-Toan Do, David Rosewarne, Gustavo Carneiro
First submitted to arxiv on: 18 Nov 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 hybrid intelligent system, Coverage-constrained Learning to Defer and Complement with Specific Experts (CL2DC), is proposed in this paper. The approach addresses the research gap in exploring both learning-to-defer (L2D) and learning-to-complement (L2C) under diverse expert knowledge to improve decision-making while controlling cooperation cost. CL2DC makes decisions through AI prediction alone, deferring to a specific expert, or complementing an expert depending on input data. A coverage-constrained optimization is also proposed to ensure the cooperation cost approximates a target probability for AI-only selection. The approach can handle scenarios with multiple noisy-label annotations without clean-label references. Comprehensive evaluations on synthetic and real-world datasets demonstrate superior performance compared to state-of-the-art HAI-CC methods. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Humans and Artificial Intelligence (AI) can work together better by sharing their strengths. Right now, AI is good at processing lots of information quickly, but humans are still better at making decisions that need creativity and common sense. In this paper, the researchers developed a new way for AI and humans to work together called Coverage-constrained Learning to Defer and Complement with Specific Experts (CL2DC). This method helps decide when to let AI make a decision alone or when to ask a human expert for help. It also takes into account the cost of asking a human expert, so it can stay within a budget. The new approach can handle situations where there are mistakes in the data and no clean references to correct them. The results show that this method performs better than other similar approaches. |
Keywords
» Artificial intelligence » Optimization » Probability
