BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Sphere Neural-Networks for Rational Reasoning II - Dr Tiansi Dong\
 , Fraunhofer Institute IAIS\, Germany
DTSTART:20241009T133000Z
DTEND:20241009T143000Z
UID:TALK222805@talks.cam.ac.uk
CONTACT:Challenger Mishra
DESCRIPTION:The success of Large Language Models (LLMs)\, e.g.\, ChatGPT\,
  is witnessed by their planetary popularity\, their capability of human-li
 ke communication\, and also by their steadily improved reasoning performan
 ce. However\, it remains unclear whether LLMs reason. It is an open proble
 m how traditional neural networks can be qualitatively extended to go beyo
 nd the statistic paradigm and achieve high-level cognition. Here\, we pres
 ent a novel qualitative extension by generalising computational building b
 locks from vectors to spheres. We propose Sphere Neural Networks (SphNNs) 
 for human-like reasoning through model construction and inspection\, and d
 evelop SphNN for syllogistic reasoning\, a microcosm of human rationality.
  SphNN is a hierarchical neuro-symbolic Kolmogorov-Arnold geometric GNN\, 
 and uses a neuro-symbolic transition map of neighbourhood spatial relation
 s to transform the current sphere configuration towards the target. SphNN 
 is the first neural model that can determine the validity of long-chained 
 syllogistic reasoning in one epoch without training data\, with the worst 
 computational complexity of O(N). SphNN can evolve into various types of r
 easoning\, such as spatio-temporal reasoning\, logical reasoning with nega
 tion and disjunction\, event reasoning\, neuro-symbolic unification\, and 
 humour understanding (the highest level of cognition). All these suggest a
  new kind of Herbert A. Simon's scissors with two neural blades. SphNNs wi
 ll tremendously enhance interdisciplinary collaborations to develop the tw
 o neural blades and realise deterministic neural reasoning and human-bound
 ed rationality and elevate LLMs to reliable psychological AI. This work su
 ggests that the non-zero radii of spheres are the missing components that 
 prevent traditional deep-learning systems from reaching the realm of ratio
 nal reasoning and cause LLMs to be trapped in the swamp of hallucination.
LOCATION:SW00\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR