4 papers
WaterSIC: Information-Theoretically (Near) Optimal Linear Layer Quantization
Egor Lifar, Semyon Savkin, Or Ordentlich +1
This paper considers the problem of converting a given dense linear layer to low precision. The tradeoff between compressed length and output discrepancy is analyzed information th…
Diffusion Domain Expansion: Learning to Coordinate Pre-trained Diffusion Models
Egor Lifar, Semyon Savkin, Timur Garipov +2
In this paper, we propose Diffusion Domain Expansion (DDE), a method that efficiently extends pre-trained diffusion models to generate larger objects and handle more complex condit…
The Radio-Frequency Transformer for Signal Separation
Egor Lifar, Semyon Savkin, Rachana Madhukara +3
We study a problem of signal separation: estimating a signal of interest (SOI) contaminated by an unknown non-Gaussian background/interference. Given the training data consisting o…
NestQuant: Nested Lattice Quantization for Matrix Products and LLMs
Semyon Savkin, Eitan Porat, Or Ordentlich +1
Post-training quantization (PTQ) has emerged as a critical technique for efficient deployment of large language models (LLMs). This work proposes NestQuant, a novel PTQ scheme for…