Search papers, labs, and topics across Lattice.
4
0
5
Aligning ternary weight quantization directly with SIMD ciphertext packing layouts eliminates up to 80% of expensive FHE multiplications, unlocking sub-1% accuracy loss speedups across both ViTs and BERT.
Physical-layer edge communication can be optimized directly for downstream LLM reasoning utility rather than raw signal-to-noise ratio, using simulated adversarial exams to score distributed sensor memories.
StateLens is presented, a framework that employs Large Language Models (LLM) to automate the discovery of deep internal states in JS engines and significantly outperforms state-of-the-art fuzzers and uncovering 68 new bugs.
