MTL for dual-output L2 ASR degrades surface transcription due to encoder-level representational entanglement, with stronger effects in English linked to surface-meaning divergence.
arXiv preprint arXiv:2104.02724 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
LARM enables test-time compute scaling in non-autoregressive ASR via depth-conditioned looping with CTC checkpoints, supervision embeddings, FiLM conditioning, and posterior feedback, yielding lower WER on LibriSpeech with more loops.
citing papers explorer
-
Multi-task Learning is Not Enough: Representational Entanglement in Dual-output Second Language Speech Recognition
MTL for dual-output L2 ASR degrades surface transcription due to encoder-level representational entanglement, with stronger effects in English linked to surface-meaning divergence.
-
Test-Time Compute Scaling for ASR with Depth-Conditioned Looped Transformers
LARM enables test-time compute scaling in non-autoregressive ASR via depth-conditioned looping with CTC checkpoints, supervision embeddings, FiLM conditioning, and posterior feedback, yielding lower WER on LibriSpeech with more loops.