“Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks
Summary: Current methodologies to model connectivity in complex networks either rely on network scientists’ intelligence to discover reliable physical rules or use artificial intelligence (AI) that stacks hundreds of inaccurate human-made rules to make a new one that optimally summarizes them togeth...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-01-01
|
| Series: | iScience |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2589004222019708 |
| _version_ | 1828057132738543616 |
|---|---|
| author | Alessandro Muscoloni Carlo Vittorio Cannistraci |
| author_facet | Alessandro Muscoloni Carlo Vittorio Cannistraci |
| author_sort | Alessandro Muscoloni |
| collection | DOAJ |
| description | Summary: Current methodologies to model connectivity in complex networks either rely on network scientists’ intelligence to discover reliable physical rules or use artificial intelligence (AI) that stacks hundreds of inaccurate human-made rules to make a new one that optimally summarizes them together. Here, we provide an accurate and reproducible scientific analysis showing that, contrary to the current belief, stacking more good link prediction rules does not necessarily improve the link prediction performance to nearly optimal as suggested by recent studies. Finally, under the light of our novel results, we discuss the pros and cons of each current state-of-the-art link prediction strategy, concluding that none of the current solutions are what the future might hold for us. Future solutions might require the design and development of next generation “creative” AI that are able to generate and understand complex physical rules for us. |
| first_indexed | 2024-04-10T21:06:30Z |
| format | Article |
| id | doaj.art-6c2f5b043fb94867895b528747def6b3 |
| institution | Directory Open Access Journal |
| issn | 2589-0042 |
| language | English |
| last_indexed | 2024-04-10T21:06:30Z |
| publishDate | 2023-01-01 |
| publisher | Elsevier |
| record_format | Article |
| series | iScience |
| spelling | doaj.art-6c2f5b043fb94867895b528747def6b32023-01-22T04:40:28ZengElsevieriScience2589-00422023-01-01261105697“Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networksAlessandro Muscoloni0Carlo Vittorio Cannistraci1Center for Complex Network Intelligence (CCNI), Tsinghua Laboratory of Brain and Intelligence (THBI), Tsinghua University, 160 Chengfu Road, SanCaiTang Building, Haidian District, Beijing 100084, ChinaCenter for Complex Network Intelligence (CCNI), Tsinghua Laboratory of Brain and Intelligence (THBI), Tsinghua University, 160 Chengfu Road, SanCaiTang Building, Haidian District, Beijing 100084, China; Department of Computer Science, Tsinghua University, Beijing, China; Department of Biomedical Engineering, Tsinghua University, Beijing, China; Corresponding authorSummary: Current methodologies to model connectivity in complex networks either rely on network scientists’ intelligence to discover reliable physical rules or use artificial intelligence (AI) that stacks hundreds of inaccurate human-made rules to make a new one that optimally summarizes them together. Here, we provide an accurate and reproducible scientific analysis showing that, contrary to the current belief, stacking more good link prediction rules does not necessarily improve the link prediction performance to nearly optimal as suggested by recent studies. Finally, under the light of our novel results, we discuss the pros and cons of each current state-of-the-art link prediction strategy, concluding that none of the current solutions are what the future might hold for us. Future solutions might require the design and development of next generation “creative” AI that are able to generate and understand complex physical rules for us.http://www.sciencedirect.com/science/article/pii/S2589004222019708Artificial intelligenceNetwork |
| spellingShingle | Alessandro Muscoloni Carlo Vittorio Cannistraci “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks iScience Artificial intelligence Network |
| title | “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks |
| title_full | “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks |
| title_fullStr | “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks |
| title_full_unstemmed | “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks |
| title_short | “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks |
| title_sort | stealing fire or stacking knowledge by machine intelligence to model link prediction in complex networks |
| topic | Artificial intelligence Network |
| url | http://www.sciencedirect.com/science/article/pii/S2589004222019708 |
| work_keys_str_mv | AT alessandromuscoloni stealingfireorstackingknowledgebymachineintelligencetomodellinkpredictionincomplexnetworks AT carlovittoriocannistraci stealingfireorstackingknowledgebymachineintelligencetomodellinkpredictionincomplexnetworks |