Joint embedding architecture
A neural network approach where inputs are encoded to embeddings and the system learns by comparing embeddings of related inputs, without reconstructing or generating original data
As defined in Yann LeCun's $1B Bet Against LLMs [Part 1]
Where it comes up · 1
Most important claims · 5 of 5Search all claims for “Joint embedding architecture” →
- Representation collapse in joint embedding architectures—where networks trivially output identical embeddings for all inputs—was the primary unsolved problem preventing joint embeddings from learning useful representations until Barlow Twins and related approaches.Yann LeCun's $1B Bet Against LLMs [Part 1]
- Horace Barlow hypothesized in 1961 that neurons in animal and human vision systems operate by reducing redundant information between neurons, a principle later applied by Stefan Deny and colleagues to solve representation collapse in joint embedding architectures.Yann LeCun's $1B Bet Against LLMs [Part 1]
- DINO V3, released August 2025, achieved 88.4% ImageNet accuracy using joint embedding architecture, marking the first time self-supervised vision models matched state-of-the-art weakly-supervised and supervised models on image classification.Yann LeCun's $1B Bet Against LLMs [Part 1]
- Siamese networks, developed by LeCun and colleagues at Bell Labs in the early 1990s for detecting fraudulent signatures, used joint embedding architecture to output similar embedding vectors for authentic signatures and different vectors for fraudulent ones, avoiding the need for generative reconstruction.Yann LeCun's $1B Bet Against LLMs [Part 1]
- Barlow Twins worked surprisingly well, avoiding representation collapse while learning powerful internal representations of images, achieving an ImageNet accuracy of 73.2% on a frozen encoder with a linear probe, outperforming the original fully supervised AlexNet model by over 10 percentage points.Yann LeCun's $1B Bet Against LLMs [Part 1]