news

Jun 26, 2024 New preprint on simultaneous translation with LLMs! SimulMask represents a significant departure from typical causal masking to unify fine-tuning and inference context management for simultaneous tasks.
Mar 16, 2024 Two papers accepted in short order! LeaPformers and Simul-LLM at ICML ‘24 and ACL ‘24 respectively. To be presented as posters at both conferences.
Jan 20, 2024 Started working with researchers at Pacific Northwest National Laboratory in a more formal fashion! Playing around with HLS and Torch-to-Verilog flow optimization with MLIR for AI acceleration.