BEGIN:VCALENDAR
VERSION:2.0
PRODID:https://evenements.uqam.ca
X-PUBLISHED-TTL:P1W
BEGIN:VEVENT
UID:30026@https://evenements.uqam.ca
DTSTART:20250220T103000Z
SEQUENCE:6
TRANSP:OPAQUE
URL:https://evenements.uqam.ca/evenements/seminaire-au-dic-theories-of-arti
 ficial-intelligence-par-herbert-roitblat/30026?date=2025-02-20_10-30-00
LOCATION:UQAM - Pavillon Président-Kennedy (PK) (201\, avenue du Présiden
 t-Kennedy\, Montréal )
SUMMARY:Séminaire au DIC: «Theories of Artificial Intelligence» par Herb
 ert Roitblat
CLASS:PUBLIC
DESCRIPTION:Séminaire ayant lieu dans le cadre du doctorat en informatique
  cognitive\, en collaboration avec le centre de recherche CRIA et l'ISC 
    \n\n\n \n\n\nHerbert ROITBLAT\n\n\nJeudi le 20 février 2025 à 10h
 30\n\n\nLocal: PK-5115   (Il est possible d'y assister en virtuel en vo
 us inscrivant ici)   \n\n\n \n\n\nTITRE :   Theories of Artificial I
 ntelligence\n\n\n \n\n\nRÉSUMÉ\n\n\nGenAI models are computationally co
 mplex\, but conceptually simple.  They are trained to fill in the blanks.
   Model semantics is limited to word distribution patterns (Harris\, 1956
 )\, yet many claim that GenAI models are capable of deep cognitive process
 es (such as reasoning and understanding). These assertions imply that cogn
 itive processes can spontaneously emerge from these behavioral patterns—
 that a theory of cognition can be constructed at the purely behavioral lev
 el of word use patterns.  We have seen that movie before\, but Chomsky (1
 959/1967) and a good deal of research with humans and animals\, peaking in
  the 1980s\, have demonstrated that a purely behavioral theory of cognitio
 n is not viable.  Those same research methods could be applied to the ana
 lysis of the latest forms of artificial intelligence\, but their relevance
  is rarely recognized.  Instead\, much of what passes for theoretical ana
 lysis of GenAI models is based on the logical fallacy of “affirming the 
 consequent.” The models behave as if they had underlying cognitive proce
 sses\, but their proponents fail to consider whether other explanations (e
 .g.\, stochastically parroting training data) could also explain the obser
 vations.  I will discuss the internal structure of GenAI models and how t
 o understand them.  I will also offer some theoretical suggestions for ap
 proaching understanding and artificial general intelligence.\n\n\n \n\n\n
 BIOGRAPHIE\n\n\nHerbert ROITBLAT\, is lead data scientist for Egnyte’s r
 esearch and development in artificial intelligence. Formerly professor of 
 Psychology\, Marine Biology\, and Second Language Acquisition\, University
  of Hawaii\, Roitblat’s work on how dolphins recognize targets underwate
 r with biosonar led to significant contributions to early neural network r
 esearch and a patent on a binaural sonar. His more recent work is on artif
 icial intelligence: “Algorithms Are Not Enough: Creating General Artific
 ial Intelligence” (MIT Press\, 2020) argues that algorithms and neural n
 etwork models cannot fully capture the complexity of human cognition or an
 imal intelligence and suggests what is needed to achieve artificial genera
 l intelligence.  \n\n\n \n\n\nRÉFÉRENCES\n\n\nBender\, E. M.\, et al.
  (2021) On the dangers of stochastic parrots: Can language models be too b
 ig? In Proceedings of the 2021 ACM. 610-623.\n\n\nChomsky\, N. (1967) A R
 eview of B. F. Skinner’s Verbal Behavior. 142-143.\n\n\nHarris\, Z. (195
 4). Distributional structure. Word\, 10(23): 146-162.\n\n\nHuang K.\, &amp
 \; Chang\, J.-C. (2023) Towards Reasoning in Large Language Models: A Surv
 ey. Findings of the Association for Computational Linguistics: ACL 2023\, 
 pages 1049–1065.\n\n\nRoitblat\, H. L. (2024). An Essay concerning machi
 ne understanding. arXiv preprint arXiv:2405.01840.\n\n\nRoitblat\, H. L. (
 2020). Algorithms are not enough: Creating general artificial intelligence
 . Mit Press.\n\n\nRoitblat\, H. L. (2017). Animal cognition. In: Bechtel 
 &amp\; Graham\, eds\, A Companion to Cognitive Science\, Wiley.\n\nMot-cl
 és : Cognition humaine\, LLMs\, LATECE\, LATECE UQAM INFORMATIQUE\, CRIA\
 , Institut des sciences cognitives\, Neurosciences\, neurosciences cogniti
 ves\, Sciences cognitives\, Apprentissage du langage naturel\, sciences du
  langage\, apprentissage machine\, apprentissage automatique\, apprentissa
 ge profond\, langage automatique\, langage cognitif\, Cognition\, intellig
 ence artificielle\, IA\, intelligence artificielle\, ChatGPT\, IA\, intell
 igence artificielle\, chatGPT\, enseignement supérieur\, IA\; intelligenc
 e artificielle\; société\, Faculté des sciences humaines\, Faculté des
  sciences de l'UQAM\, Faculté des sciences\, département de psychologie\
 , Département d'informatique\, doctorat en psychologie\, doctorat en info
 rmatique\, doctorat en informatique cognitive\n\nPrix : Gratuit\n\n
CATEGORIES:Séminaire,Conférence
DTSTAMP:20260914T054358Z
CREATED:20250115T122038Z
LAST-MODIFIED:20250115T141956Z
END:VEVENT
END:VCALENDAR