π-Gen (PAI-Gen) project

π-Gen = PAI-Gen = Profound AI Genius

Aim of project: disruptive innovations to reshape AI

Rule-Augmented Learning (RAL)

apply existing induced complicated rules (like grammar) with (domain/partial) knowledge to drastically boost learning efficiency, including:

  • GID (German equipped Infantry Division of China) ensemble model: (prototype constructed)
  • InA algorithm: for language comprehension (in progress)

Inductive Learning (IL)

Learn by inducing new general rules, which will be the fourth learning method after Supervised Learning, Unsupervised Learning, Reinforcement Learning. IL is generalisable, explainable, transferable, adaptable to small datasets, including:

  • DoNde algorithm: for tabular data (near finishing)
  • ViC-SeIL algorithm: for image recognition (in progress)
  • NeGSt algorithm: AI4Science (starting)
  • AdaStra algorithm

Analogical Learning (AL)

learn from indirectly relevant and similar data, aka cross domain generalisation. AL will be the fifth learning method

Emotional AI (EAI)

Emotional Game Theory

Scroll to Top