REVIEW 4 cited by
Is Cosine-Similarity of Embeddings Really About Similarity?
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Is Cosine-Similarity of Embeddings Really About Similarity?
read the original abstract
Cosine-similarity is the cosine of the angle between two vectors, or equivalently the dot product between their normalizations. A popular application is to quantify semantic similarity between high-dimensional objects by applying cosine-similarity to a learned low-dimensional feature embedding. This can work better but sometimes also worse than the unnormalized dot-product between embedded vectors in practice. To gain insight into this empirical observation, we study embeddings derived from regularized linear models, where closed-form solutions facilitate analytical insights. We derive analytically how cosine-similarity can yield arbitrary and therefore meaningless `similarities.' For some linear models the similarities are not even unique, while for others they are implicitly controlled by the regularization. We discuss implications beyond linear models: a combination of different regularizations are employed when learning deep models; these have implicit and unintended effects when taking cosine-similarities of the resulting embeddings, rendering results opaque and possibly arbitrary. Based on these insights, we caution against blindly using cosine-similarity and outline alternatives.
Forward citations
Cited by 4 Pith papers
-
Matter to Mechanism: A Benchmark for AI Co-Scientists in Materials and Battery Research
Introduces the Matter to Mechanism benchmark of 2,645 structured instances and a composite metric suite for evaluating AI co-scientists on problem-to-hypothesis reasoning in battery materials research.
-
GENIUS: An Agentic AI Framework for Autonomous Design and Execution of Simulation Protocols
GENIUS is an agentic AI framework that automates generation, validation, and repair of Quantum ESPRESSO DFT input files, succeeding on ~80% of 295 benchmarks with 76% autonomous repairs and lower cost than LLM-only baselines.
-
How Small Transformation Expose the Weakness of Semantic Similarity Measures
A diagnostic benchmark of text and code transformations finds embedding similarity metrics often conflate opposition with equivalence; LLM judges discriminate better, and Euclidean distance improves code embeddings.
-
In Defense of Cosine Similarity: Normalization Eliminates the Gauge Freedom
On unit-normalized embeddings, cosine distance equals half the squared Euclidean distance, so the diagonal gauge ambiguity vanishes when normalization is imposed during training.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.