Alert button
Picture for Anshumali Shrivastava

Anshumali Shrivastava

Alert button

Distributed SLIDE: Enabling Training Large Neural Networks on Low Bandwidth and Simple CPU-Clusters via Model Parallelism and Sparsity

Jan 29, 2022
Minghao Yan, Nicholas Meisburger, Tharun Medini, Anshumali Shrivastava

Viaarxiv icon

Breaking the Linear Iteration Cost Barrier for Some Well-known Conditional Gradient Methods Using MaxIP Data-structures

Nov 30, 2021
Anshumali Shrivastava, Zhao Song, Zhaozhuo Xu

Figure 1 for Breaking the Linear Iteration Cost Barrier for Some Well-known Conditional Gradient Methods Using MaxIP Data-structures
Figure 2 for Breaking the Linear Iteration Cost Barrier for Some Well-known Conditional Gradient Methods Using MaxIP Data-structures
Figure 3 for Breaking the Linear Iteration Cost Barrier for Some Well-known Conditional Gradient Methods Using MaxIP Data-structures
Figure 4 for Breaking the Linear Iteration Cost Barrier for Some Well-known Conditional Gradient Methods Using MaxIP Data-structures
Viaarxiv icon

Federated Multiple Label Hashing (FedMLH): Communication Efficient Federated Learning on Extreme Classification Tasks

Oct 23, 2021
Zhenwei Dai, Chen Dun, Yuxin Tang, Anastasios Kyrillidis, Anshumali Shrivastava

Figure 1 for Federated Multiple Label Hashing (FedMLH): Communication Efficient Federated Learning on Extreme Classification Tasks
Figure 2 for Federated Multiple Label Hashing (FedMLH): Communication Efficient Federated Learning on Extreme Classification Tasks
Figure 3 for Federated Multiple Label Hashing (FedMLH): Communication Efficient Federated Learning on Extreme Classification Tasks
Figure 4 for Federated Multiple Label Hashing (FedMLH): Communication Efficient Federated Learning on Extreme Classification Tasks
Viaarxiv icon

Random Offset Block Embedding Array (ROBE) for CriteoTB Benchmark MLPerf DLRM Model : 1000$\times$ Compression and 2.7$\times$ Faster Inference

Aug 04, 2021
Aditya Desai, Li Chou, Anshumali Shrivastava

Figure 1 for Random Offset Block Embedding Array (ROBE) for CriteoTB Benchmark MLPerf DLRM Model : 1000$\times$ Compression and 2.7$\times$ Faster Inference
Figure 2 for Random Offset Block Embedding Array (ROBE) for CriteoTB Benchmark MLPerf DLRM Model : 1000$\times$ Compression and 2.7$\times$ Faster Inference
Figure 3 for Random Offset Block Embedding Array (ROBE) for CriteoTB Benchmark MLPerf DLRM Model : 1000$\times$ Compression and 2.7$\times$ Faster Inference
Figure 4 for Random Offset Block Embedding Array (ROBE) for CriteoTB Benchmark MLPerf DLRM Model : 1000$\times$ Compression and 2.7$\times$ Faster Inference
Viaarxiv icon

Efficient Inference via Universal LSH Kernel

Jun 21, 2021
Zichang Liu, Benjamin Coleman, Anshumali Shrivastava

Figure 1 for Efficient Inference via Universal LSH Kernel
Figure 2 for Efficient Inference via Universal LSH Kernel
Figure 3 for Efficient Inference via Universal LSH Kernel
Figure 4 for Efficient Inference via Universal LSH Kernel
Viaarxiv icon

PairConnect: A Compute-Efficient MLP Alternative to Attention

Jun 15, 2021
Zhaozhuo Xu, Minghao Yan, Junyan Zhang, Anshumali Shrivastava

Figure 1 for PairConnect: A Compute-Efficient MLP Alternative to Attention
Figure 2 for PairConnect: A Compute-Efficient MLP Alternative to Attention
Figure 3 for PairConnect: A Compute-Efficient MLP Alternative to Attention
Figure 4 for PairConnect: A Compute-Efficient MLP Alternative to Attention
Viaarxiv icon

Sublinear Least-Squares Value Iteration via Locality Sensitive Hashing

Jun 08, 2021
Anshumali Shrivastava, Zhao Song, Zhaozhuo Xu

Figure 1 for Sublinear Least-Squares Value Iteration via Locality Sensitive Hashing
Figure 2 for Sublinear Least-Squares Value Iteration via Locality Sensitive Hashing
Figure 3 for Sublinear Least-Squares Value Iteration via Locality Sensitive Hashing
Figure 4 for Sublinear Least-Squares Value Iteration via Locality Sensitive Hashing
Viaarxiv icon

IRLI: Iterative Re-partitioning for Learning to Index

Mar 17, 2021
Gaurav Gupta, Tharun Medini, Anshumali Shrivastava, Alexander J Smola

Figure 1 for IRLI: Iterative Re-partitioning for Learning to Index
Figure 2 for IRLI: Iterative Re-partitioning for Learning to Index
Figure 3 for IRLI: Iterative Re-partitioning for Learning to Index
Figure 4 for IRLI: Iterative Re-partitioning for Learning to Index
Viaarxiv icon

Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More

Mar 06, 2021
Shabnam Daghaghi, Nicholas Meisburger, Mengnan Zhao, Yong Wu, Sameh Gobriel, Charlie Tai, Anshumali Shrivastava

Figure 1 for Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More
Figure 2 for Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More
Figure 3 for Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More
Figure 4 for Accelerating SLIDE Deep Learning on Modern CPUs: Vectorization, Quantizations, Memory Optimizations, and More
Viaarxiv icon