Meta's I-JEPA: First AI model based on LeCun’s vision for more human-like AI

I&#x27;ve been incredibly happy with Meta&#x27;s decision to open source so much of their work. When US Senators Josh Hawley and Richard Blumenthal attacked Meta for open sourcing LLaMA I was worried they would retreat.<p>Yann LeCun has greatly increased my understanding of AI by being so open with his work and thoughts.<p>(<a href="https:&#x2F;&#x2F;www.hawley.senate.gov&#x2F;hawley-and-blumenthal-demand-answers-meta-warn-misuse-after-leak-metas-ai-model" rel="nofollow noreferrer">https:&#x2F;&#x2F;www.hawley.senate.gov&#x2F;hawley-and-blumenthal-demand-a...</a>)

From the blog:<p>&quot;Our paper on I-JEPA (<a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2301.08243" rel="nofollow noreferrer">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2301.08243</a>) will be presented at CVPR 2023 next week, and we’re also open-sourcing the training code and model checkpoints today.&quot;