LightBeam: An Accurate and Memory-Efficient CTC Decoder for Speech Neuroprostheses

TL;DR

LightBeam is a memory-efficient, non-WFST CTC decoder that achieves state-of-the-art speech decoding performance from neural data with significantly reduced RAM requirements, enabling broader accessibility.

Contribution

The paper introduces LightBeam, a novel CTC decoder that reduces memory usage from 320 GB to 10 GB while maintaining top performance, by integrating an LLM into the beam search.

Findings

Achieves state-of-the-art results on Brain-to-Text benchmarks.

Reduces RAM requirement from 320 GB to 10 GB.

Open-source implementation in Python.

Abstract

A promising pathway for restoring communication in patients with dysarthria and anarthria is speech neuroprostheses, which directly decode speech from cortical neural activity. Two benchmarks, Brain-to-Text '24 and '25, released intracranial recordings from patients with dysarthria along with a baseline algorithm trained with Connectionist Temporal Classification (CTC). Despite significant innovation on these benchmarks, all leading published prior work relies on a WFST-based CTC decoder that requires ${\sim}$320 GB of RAM. These memory requirements limit accessibility for both patients and researchers. Here, we propose LightBeam, a non-WFST based CTC decoder that requires only ${\sim}$10 GB of RAM and achieves state-of-the-art performance on both benchmarks. LightBeam achieves this by integrating an LLM into the beam-search process via delayed fusion, obviating the prior need for using a large N-gram LM. LightBeam is implemented in Python and is open-source.