When do compressed vectors make the same decisions?
Places quantization error inside the ranking and graph-pruning decisions that vector search actually makes, relating reliability to comparison margins, correlated residuals, and execution traces.
RESEARCH / PUBLICATIONS
Questions.
Evidence. Progress.
Our research begins with questions worth asking, then builds methods and evidence across different fields. This page brings together papers, current directions, and the next questions opened by each result.
Why does the value of the same quantization move change?
Studies the real utility of a discrete move in extreme low-bit LLMs by evaluating the move along its own path and combining decisions in an evolving model state.
What survives at one or two bits?
Explains low-bit ranking behavior through covariance structure and coordinate heterogeneity, including why an extra magnitude bit and random rotation help different representations in different ways.
QuIVer: building the search graph directly in binary space
Builds, prunes, and navigates an ANN graph in a training-free two-bit space, reading full vectors only for final reranking and exposing the data regimes where compact topology works.
No matching paper was found. Try another keyword or research direction.
4 papers
EXPLORE THE DIRECTIONS
Follow a question into its field.
Browse the foundations that connect our work in data systems, retrieval, representation, efficient intelligence, and mathematics.
Have a question worth pursuing together?
Talk with us