LLC: Accurate, Multi-purpose Learnt Low-dimensional Binary Codes

Abstract

Learning binary representations of instances and classes is a classical problem with several high potential applications. In modern settings, the compression of high-dimensional neural representations to low-dimensional binary codes is a challenging task and often require large bit-codes to be accurate. In this work, we propose a novel method for Learning Low-dimensional binary Codes (LLC) for instances as well as classes. Our method does not require any side-information, like annotated attributes or label meta-data, and learns extremely low-dimensional binary codes (~20 bits for ImageNet-1K). The learnt codes are super-efficient while still ensuring nearly optimal classification accuracy for ResNet50 on ImageNet-1K. We demonstrate that the learnt codes capture intrinsically important features in the data, by discovering an intuitive taxonomy over classes. We further quantitatively measure the quality of our codes by applying it to the efficient image retrieval as well as out-of-distri

Stats

• citations0
• S2 citations—
• github stars13
• HF likes0
• heat score2.29
• arxiv keykusupati2021llc

Related papers

• Optimized Feature Space Learning For Generating Efficient Binary Codes For Image Retrieval (2020)5.24
• Compact Hash Code Learning With Binary Deep Neural Network (2017)9.03
• Unsupervised Representation Learning Via Neural Activation Coding (2021)1.56
• Binary Constrained Deep Hashing Network For Image Retrieval Without Manual Annotation (2018)5.84
• End-to-end Binary Representation Learning Via Direct Binary Embedding (2017)5.84
• Embarrassingly Simple Binary Representation Learning (2019)9.41
• GLARE: Low Light Image Enhancement Via Generative Latent Feature Based Codebook Retrieval (2024)16.37
• A Multi-layer Network Based On Sparse Ternary Codes For Universal Vector Compression (2017)0.00