In-memory Nearest Neighbor Search With Fefet Multi-bit Content-addressable Memories

Abstract

Nearest neighbor (NN) search is an essential operation in many applications, such as one/few-shot learning and image classification. As such, fast and low-energy hardware support for accurate NN search is highly desirable. Ternary content-addressable memories (TCAMs) have been proposed to accelerate NN search for few-shot learning tasks by implementing \(L_\infty\) and Hamming distance metrics, but they cannot achieve software-comparable accuracies. This paper proposes a novel distance function that can be natively evaluated with multi-bit content-addressable memories (MCAMs) based on ferroelectric FETs (FeFETs) to perform a single-step, in-memory NN search. Moreover, this approach achieves accuracies comparable to floating-point precision implementations in software for NN classification and one/few-shot learning tasks. As an example, the proposed method achieves a 98.34% accuracy for a 5-way, 5-shot classification task for the Omniglot dataset (only 0.8% lower than software-based imp

Stats

• citations37
• S2 citations—
• github stars0
• HF likes0
• heat score11.85
• arxiv keykazemi2020in

Related papers

• Efficient And Reliable Vector Similarity Search Using Asymmetric Encoding With Nand-flash For Many-class Few-shot Learning (2024)2.26
• An Empirical Comparison Of FAISS And FENSHSES For Nearest Neighbor Search In Hamming Space (2019)0.00
• Associative Memories To Accelerate Approximate Nearest Neighbor Search (2016)6.34
• Fast And Exact Nearest Neighbor Search In Hamming Space On Full-text Search Engines (2019)5.24
• PM-LSH: A Fast And Accurate In-memory Framework For High-dimensional Approximate NN And Closest Pair Search (2021)8.09
• Federated Nearest Neighbor Classification With A Colony Of Fruit-flies: With Supplement (2021)0.00
• Neuromorphic Computing For Content-based Image Retrieval (2020)9.03
• Fast And Bayes-consistent Nearest Neighbors (2019)0.00