Description du projet
Stimuler les découvertes scientifiques en automatisant la recherche d’analogies
Les analogies sont souvent utilisées dans le domaine scientifique. On leur doit de nombreuses découvertes importantes dans l’histoire. Le raisonnement analogique repose sur la capacité à identifier une règle structurelle profonde et à l’appliquer à différents domaines. Aujourd’hui, avec l’aide de l’IA et de l’apprentissage automatique, la quantité de données et la multiplicité des modèles structurels disponibles augmentent considérablement la possibilité de faire des découvertes. Le projet SIAM, financé par l’UE, entend mettre au point un mécanisme d’automatisation du processus de recherche d’analogies en combinant l’innovation humaine et le traitement automatisé de l’information. Il utilisera l’IA pour l’identification, la sélection et l’application des analogies. Le projet préparera des outils pour comparer les analogies et détecter des similitudes, en vue de créer des algorithmes reposant sur le bon sens et l’abstraction pour développer de nouveaux instruments qui accéléreront l’innovation et la découverte.
Objectif
"Many world-changing breakthroughs in science and technology were enabled by analogical transfer, as ideas from one domain were used to solve a problem in another. Observing water led the Greek philosopher Chrysippus to speculate that sound was a wave phenomenon; an analogy to twisting a cardboard box allowed the Wright brothers to design a steerable aircraft. Despite its value for innovation, very little progress has been made towards automating the process of analogy-finding in real-world settings, and the problem has maintained a longstanding status as a ""holy grail"" in artificial intelligence (AI).
The goal of this proposal is to tackle head-on this important problem and develop principled tools for automatically discovering analogies in large, unstructured, natural-language datasets such as patents and scientific papers. Such tools could revolutionize a variety of fields, allowing scientists and inventors to retrieve useful content based on deep structural similarity rather than simple keywords. The explosion of data available online, coupled with novel machine learning and crowdsourcing techniques, creates an unprecedented opportunity to develop novel methods to accelerate innovation and discovery.
My approach explores the multiple roles AI and machine learning can play in the analogical innovation pipeline. This research will focus on the three core components of the pipeline -- (1) developing representations and similarity metrics to facilitate comparison between potential analogs, (2) imbuing the algorithms with commonsense knowledge and abstraction capabilities, and (3) guiding the adaptation of the discovered analogies to solve the original problem. For each component, the proposal demonstrates how recent advances suggest effective approaches, and describes our concrete preliminary results and ideas to serve as starting points and indicate the feasibility of this challenging project."
Champ scientifique
Not validated
Not validated
Mots‑clés
Programme(s)
Thème(s)
Régime de financement
ERC-STG - Starting GrantInstitution d’accueil
91904 Jerusalem
Israël