Deep Learning Pioneers Before Hinton | Generated by AI

Home 2025.10

Alexey Ivakhnenko’s Work

Alexey Grigoryevich Ivakhnenko, a Ukrainian-Soviet cyberneticist, is often called the “godfather of deep learning” for his pioneering efforts in the 1960s and 1970s. Working under resource constraints during the Cold War era—when computing power was millions of times more limited than today—he focused on multilayer neural networks that could automatically learn hierarchical representations of data.

Ivakhnenko’s GMDH evolved into a broader inductive modeling framework, influencing fields like control systems and economics. Despite its impact, much of his work was published in Russian and overlooked in English-language AI circles.

Shun-ichi Amari’s Work

Shun-ichi Amari, a Japanese mathematician and neuroscientist, made foundational contributions to neural network theory in the 1960s and 1970s, emphasizing adaptive learning and geometric perspectives on information processing. His research bridged neuroscience and computation, laying groundwork for self-organizing systems.

Amari also founded information geometry, a field using differential geometry to analyze statistical models and neural dynamics, which underpins modern probabilistic neural networks.

Context in the 2024 Nobel Backlash

In his 2024 report “A Nobel Prize for Plagiarism,” Jürgen Schmidhuber argues that Hinton and Hopfield’s Nobel-winning ideas—such as the Boltzmann machine (1985) for learning representations and the Hopfield network (1982) for associative memory—repackaged Ivakhnenko’s layer-wise deep learning and Amari’s SGD/adaptive recurrent models without attribution. For instance, the Boltzmann machine omitted citations to Ivakhnenko’s 1965 internal representation learning and Amari’s 1967 SGD, while Hopfield’s network ignored Amari’s 1972 adaptive Ising model. Schmidhuber claims this omission persisted in later surveys and the Nobel’s scientific background document, highlighting a Western-centric bias in AI history.

Alexey Ivakhnenko - Wikipedia
Deep Learning in a Nutshell: History and Training - NVIDIA Developer
Shun-ichi Amari | Kyoto Prize
Japanese scientists were pioneers of AI, yet they’re being written out of history
A Nobel Prize for Plagiarism - Technical Report IDSIA-24-24
Machine Learning Stirs Controversy in Nobel Prize in Physics - Harvard Technology Review


Back

x-ai/grok-4-fast

Donate