AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

End-to-End Contrastive Language-Speech Pretraining Model for Long-Form Spoken Question Answering

Jiliang Hu, Zuchao Li, Baoyuan Qi, Guoming Liu, Ping Wang

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Sign and cross-lingual translation · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add VQ adaptor (quantizer) (ablation)CLSR · Spoken-SQuADQ-C R@1 (%): 15.29 -> 42.52Improved
add pre-trained BGE in joint training (ablation)CLSR · Spoken-SQuADQ-C R@1 (%): 42.52 -> 48.10Improved
add ASR (Paraformer) pre-training (ablation)CLSR · Spoken-SQuADWER (%): 17.00 -> 16.18Improved
add ASR (Paraformer) pre-training (ablation)CLSR · Spoken-SQuADQ-C R@1 (%): 48.10 -> 49.00Improved
add sampler (ablation)CLSR · Spoken-SQuADWER (%): 16.18 -> 15.01Improved
add sampler (ablation)CLSR · Spoken-SQuADQ-C R@1 (%): 49.00 -> 49.65Improved
add post-training of BGE (ablation)CLSR · Spoken-SQuADQ-C R@1 (%): 49.65 -> 49.82No clear effect
add post-training of BGE (ablation)CLSR · Spoken-SQuADQ-C R@10 (%): 85.91 -> 85.83No clear effect
use CLSR text-like representation instead of direct speech-text alignment (ParaBGE) (ablation)CLSR · Spoken-SQuADQ-C R@1 (%): 38.68 -> 49.82Improved
use CLSR instead of ParaBGECLSR · SLUE-SQA-5WER (%): 7.31 -> 16.69Hurt
add CLSR retrieval before Qwen-Audio for long audioQwen-Audio · Long-audio SQAEM: 18.00 -> 27.60Improved
add CLSR retrieval before Qwen-Audio for long audioQwen-Audio · Long-audio SQAcost time (s): 7935.00 -> 783.00Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy