Two Papers Accepted at Interspeech 2025!

Two of our papers have been accepted to the 26th edition of the Interspeech Conference, to be held on August 17–21, 2025 in Rotterdam, The Netherlands. These works explore concept formation in speech models and multilingual spoken interaction for LLMs.


🧠 From Words to Waves

Title: From Words to Waves: Analyzing Concept Formation in Speech and Text-Based Foundation Models Authors: Asim Ersoy, Basel Ahmad Mousi, Shammur Absar Chowdhury, Firoj Alam, Fahim I Dalvi, Nadir Durrani Summary: This paper investigates how foundation models in both speech and text domains form semantic concepts, offering insights into their internal representations and cross-modal generalization.


🗣️ SpokenNativQA: Multilingual Spoken Queries

Title: SpokenNativQA: Multilingual Everyday Spoken Queries for LLMs Authors: Firoj Alam, Md Arid Hasan, Shammur Absar Chowdhury Summary: We introduce SpokenNativQA, a dataset and benchmark designed to evaluate multilingual large language models on culturally grounded, spoken natural queries.





Enjoy Reading This Article?

Here are some more articles you might like to read next:

  • ArAIEval Shared Task: Propagandistic Techniques Detection in Unimodal and Multimodal Arabic Content
  • ArAIEval Shared Task: Propagandistic Techniques Detection in Unimodal and Multimodal Arabic Content
  • Can GPT-4 Identify Propaganda?
  • LAraBench: Benchmarking Arabic AI with Large Language Models