Abstract
Ten fundamental to-be-solved problems for current artificial intelligence research and development are proposed and analyzed from the mathematical and physical points of view. These 10 problems include: (1) the statistical foundation problem for big data analysis, (2) the algorithm problem for basic mathematical computation in big data circumstances, (3) the understanding of structures and properties of various data spaces, (4) the mathematical theories on deep learning, (5) the investigation on the optimal transportation problem under irregular constraints, (6) the problem of how to simulate learning methodology and learn in infinite functional spaces, (7) breaking through the prior hypotheses of machine learning, (8) realizing machine learning automation from selection to creation, (9) integrating knowledge inference and data learning, (10) intelligent search and AI chips.
| Translated title of the contribution | Ten fundamental problems for artificial intelligence: mathematical and physical aspects |
|---|---|
| Original language | Chinese (Traditional) |
| Pages (from-to) | 1967-1978 |
| Number of pages | 12 |
| Journal | Scientia Sinica Informationis |
| Volume | 51 |
| Issue number | 12 |
| DOIs | |
| State | Published - Dec 2021 |
Fingerprint
Dive into the research topics of 'Ten fundamental problems for artificial intelligence: mathematical and physical aspects'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver