AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Minimizing FLOPs to Learn Efficient Sparse Representations

Biswajit Paria, Chih-Kuan Yeh, Ian E.H. Yen, Ning Xu, Pradeep Ravikumar, Barnabás Póczos

ICLR 2020 · Architectures and efficiency › RNNs and normalization · approach: Learned embeddings for retrieval · 7 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use the FLOPs regularizer instead of L1 regularization64-dimensional sparse embeddings · CIFAR-100 testFLOPs-per-row: 0.47 -> 0.40Improved
Use the FLOPs regularizer instead of L1 regularization64-dimensional sparse embeddings · CIFAR-100 testprec@4: 54.32 -> 55.23Improved
Use the FLOPs regularizer instead of L1 regularization64-dimensional sparse embeddings · CIFAR-100 testprec@16: 54.96 -> 55.21Improved
Increase the number of active hash bits from k=1 to k=3 (ablation)SDH · CIFAR-100 testprec@4: 57.22 -> 55.81Hurt
Increase the number of active hash bits from k=1 to k=3 (ablation)SDH · CIFAR-100 testFLOPs-per-row: 1.18 -> 8.82Hurt
Use sparse FLOPs-regularized embeddings instead of dense embeddings64-dimensional embeddings · CIFAR-100 testFLOPs-per-row: 64 -> 0.40Improved
Use sparse FLOPs-regularized embeddings instead of dense embeddings64-dimensional embeddings · CIFAR-100 testprec@4: 57.31 -> 55.23Hurt
Use sparse FLOPs-regularized embeddings instead of SDH k=364-dimensional embeddings · CIFAR-100 testFLOPs-per-row: 8.82 -> 0.40Improved
Use sparse FLOPs-regularized embeddings instead of SDH k=364-dimensional embeddings · CIFAR-100 testprec@4: 55.81 -> 55.23Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy