@spatialmlnet

Center for Spatial Computational Learning https://xitter.cf/t.co/helWAckjG6

Joined May 2019
  • Tweets478
  • Following38
  • Followers242
28Media
We are proud to announce that the @EPSRC @spatialmlnet project has completed. This feed may not be maintained in the future. Thank you to all who contributed to the successes documented in the history of this feed. We have built a self-sustaining community to be proud of.
2
114
"Beyond Network Switching: FPGA-based Switch Architecture for Fast and Accurate Ensemble Learning" by @spatialmlnet member Luk and coauthors, now at preprints.org/manuscript/202…

Beyond Network Switching: FPGA-based Switch Architecture for Fast and Accurate Ensemble Learning

Recent advancements in Artificial Intelligence (AI) and generative models have significantly accelerated the evolution of internet services. In data centers, where data is often distributed across...

preprints.org
2
98
"Learning to Compare Hardware Designs for High-Level Synthesis" by @spatialmlnet members Sohrabizadeh and Cong and their coauthors, now at doi.org/10.1145/3670474.3685…

Learning to Compare Hardware Designs for High-Level Synthesis | Proceedings of the 2024 ACM/IEEE...

dl.acm.org
1
66
"Efficient Task Transfer for HLS DSE" by @spatialmlnet affiliates Sohrabizadeh and Cong and their coauthors, now at arxiv.org/abs/2408.13270
1
60
"H2PIPE: High throughput CNN Inference on FPGAs with High-Bandwidth Memory" by @spatialmlnet affiliate Betz and coauthors. Now at: arxiv.org/abs/2408.09209
2
2
118
"Highly-Parallel CNN Accelerator for RepVGG-like Network Training on FPGAs" by @spatialmlnet affiliates Boland and Leong and coauthors, now at doi.org/10.1109/TCAD.2024.34…
34
"FPGA Acceleration of Dynamic Neural Networks: Challenges and Advancements" by @spatialmlnet students Dimitriou and Bings and investigators @jon_hare and @g_merrett now at doi.org/10.1109/COINS61597.2…
66
Suhail Basalama's talk is now on YouTube at youtube.com/watch?v=CXH6Uu_g…
spatialml@spatialmlnet20 Jul 2024
Join us 9am EST (2pm UK time) on Tuesday on Zoom to hear about Suhail's work at @UCLA!
1
1
3
264
"Multi-Token Joint Speculative Decoding for Accelerating Large Language Model Inference" by @spatialmlnet affiliate Cong and coauthors, now at arxiv.org/abs/2407.09722v1
48
"Accelerating MRI Uncertainty Estimation with Mask-based Bayesian Neural Network" by @spatialmlnet affiliate Luk and coauthors. Now at arxiv.org/abs/2407.05521
75
Kratos: An FPGA Benchmark for Unrolled DNNs with Fine-Grained Sparsity and Mixed Precision by @spatialmlnet affiliates Chen and @mohsaied and coauthor. Now at arxiv.org/abs/2407.06033

Kratos: An FPGA Benchmark for Unrolled DNNs with Fine-Grained...

FPGAs offer a flexible platform for accelerating deep neural network (DNN) inference, particularly for non-uniform workloads featuring fine-grained unstructured sparsity and mixed arithmetic...

arxiv.org
1
45
"ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models" by @spatialmlnet affiliates @AmishAcorns, Zhang, @mohsaied and coauthors. Now at arxiv.org/abs/2406.16635

ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models

The high power consumption and latency-sensitive deployments of large language models (LLMs) have motivated efficiency techniques like quantization and sparsity. Contextual sparsity, where the...

arxiv.org
1
64
"Ultrafast jet classification at the HL-LHC" by @spatialmlnet affiliate Luk and coauthors, now at osti.gov/pages/biblio/238700…
39
Join us 9am EST (2pm UK time) on Tuesday on Zoom to hear about Suhail's work at @UCLA!
1
1
1
4
330
(see pinned post for Zoom link)
55
Our latest seminar is now up. Hear Camilo Vega from @UofT speak about the Galapagos multi-FPGA architecture! youtube.com/watch?v=TGeepxMb…
1
81
"ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models" by @spatialmlnet affiliates @AmishAcorns, Zhang, @mohsaied and coauthors. Now at arxiv.org/abs/2406.16635

ShadowLLM: Predictor-based Contextual Sparsity for Large Language Models

The high power consumption and latency-sensitive deployments of large language models (LLMs) have motivated efficiency techniques like quantization and sparsity. Contextual sparsity, where the...

arxiv.org
2
78
"Towards Open Respiratory Acoustic Foundation Models: Pretraining and Benchmarking" by @spatialmlnet affiliate @jag_jagmohan and coauthors. Now at arxiv.org/abs/2406.16148
43
"Unlocking the Global Synergies in Low-Rank Adapters" by @spatialmlnet affiliates Zhang, @gconstantinides, @aaronzhao123 and coauthors. Now at arxiv.org/abs/2406.14956
57
"Optimised Grouped-Query Attention Mechanism for Transformers" by @spatialmlnet affiliates Zhang, @gconstantinides, @aaronzhao123 and coauthors. Now at arxiv.org/abs/2406.14963

Optimised Grouped-Query Attention Mechanism for Transformers

Grouped-query attention (GQA) has been widely adopted in LLMs to mitigate the complexity of multi-head attention (MHA). To transform an MHA to a GQA, neighbour queries in MHA are evenly split into...

arxiv.org
1
76
Load more