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 - Dr Tiansi Dong\, F
 raunhofer Institute IAIS\, Germany
DTSTART:20241008T140000Z
DTEND:20241008T153000Z
UID:TALK222721@talks.cam.ac.uk
CONTACT:Challenger Mishra
DESCRIPTION:This is an up-to-date introduction to Sphere Neural-Network. T
 he last one can be viewed at http://www.youtube.com/watch?v=mZMwYtwymm0&t=
 1s\n\nThe success of Large Language Models (LLMs)\, e.g.\, ChatGPT\, is wi
 tnessed by their planetary popularity\, their capability of human-like com
 munication\, and also by their steadily improved reasoning performance. Ho
 wever\, it remains unclear whether LLMs reason. It is an open problem how 
 traditional neural networks can be qualitatively extended to go beyond the
  statistic paradigm and achieve high-level cognition. Here\, we present a 
 novel qualitative extension by generalising computational building blocks 
 from vectors to spheres. We propose Sphere Neural Networks (SphNNs) for hu
 man-like reasoning through model construction and inspection\, and develop
  SphNN for syllogistic reasoning\, a microcosm of human rationality. SphNN
  is a hierarchical neuro-symbolic Kolmogorov-Arnold geometric GNN\, and us
 es a neuro-symbolic transition map of neighbourhood spatial relations to t
 ransform the current sphere configuration towards the target. SphNN is the
  first neural model that can determine the validity of long-chained syllog
 istic reasoning in one epoch without training data\, with the worst comput
 ational complexity of O(N). SphNN can evolve into various types of reasoni
 ng\, such as spatio-temporal reasoning\, logical reasoning with negation a
 nd disjunction\, event reasoning\, neuro-symbolic unification\, and humour
  understanding (the highest level of cognition). All these suggest a new k
 ind of Herbert A. Simon's scissors with two neural blades. SphNNs will tre
 mendously enhance interdisciplinary collaborations to develop the two neur
 al blades and realise deterministic neural reasoning and human-bounded rat
 ionality and elevate LLMs to reliable psychological AI. This work suggests
  that the non-zero radii of spheres are the missing components that preven
 t traditional deep-learning systems from reaching the realm of rational re
 asoning and cause LLMs to be trapped in the swamp of hallucination.
LOCATION:SW00\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR