The Responsibility Problem in Neural Networks with Unordered Targets
| Title: | The Responsibility Problem in Neural Networks with Unordered Targets |
|---|---|
| Authors: | Hayes, Ben, Saitis, Charalampos, Fazekas, György |
| Publication Year: | 2023 |
| Collection: | Computer Science |
| Subject Terms: | Computer Science - Machine Learning, Computer Science - Discrete Mathematics |
| More Details: | We discuss the discontinuities that arise when mapping unordered objects to neural network outputs of fixed permutation, referred to as the responsibility problem. Prior work has proved the existence of the issue by identifying a single discontinuity. Here, we show that discontinuities under such models are uncountably infinite, motivating further research into neural networks for unordered data. Comment: Accepted for TinyPaper archival at ICLR 2023: https://openreview.net/forum?id=jd7Hy1jRiv4 |
| Document Type: | Working Paper |
| Access URL: | http://arxiv.org/abs/2304.09499 |
| Accession Number: | edsarx.2304.09499 |
| Database: | arXiv |
| Description not available. |