vec2vec-architecture-input-output-adapters-shared-backbone

IN premise — summaries/2026/08/24/jha-2025-vec2vec-s2-problem-formulation-unsupervised-embedding-translation.md

Created 2026-08-24T17:10:57+00:00

vec2vec uses a modular architecture with space-specific input adapters (A₁, A₂), a shared backbone (T), and output adapters (B₁, B₂), where translation is F₁ = B₂ ∘ T ∘ A₁ and reconstruction is R₁ = B₁ ∘ T ∘ A₁.

Summary

The heavy computational work is done once in a single shared core, and you only swap out the lightweight "plugs" at each end to decide what the system does with the result. This means adding a new task or space doesn't require retraining the whole model — you just attach a new adapter pair and the shared backbone handles the rest.