Our #PickOfTheWeek by @mgaido91: "AZEROS: Extending LLM to Speech with Self-Generated Instruction-Free Tuning" by Yiwen Shao, Wei Liu, Jiahong Li, Tianzi Wang, Kun Wei, Meng Yu, Dong Yu (2025)
#Speech #SpeechTech
cool work showing we can build generic speechLLM capable of addressing a wide range of tasks by leveraging only ASR data and training a small adapter between the speech encoder and LLM! https://arxiv.org/pdf/2601.06086
@fbk_mt
The Meetween consortium met in Rome for a two-day in-person session hosted by @Translation and @picampusschool.
Partners aligned on SpeechLMM progress, roadmap, ethics and next steps toward open, trustworthy speech foundation models.
#Meetween #EuropeanAI #OpenSourceAI
Announcing the Safeguards Challenge, the Arena's first Red-versus-Blue competition.
$140K in prizes. Red teams attack, blue teams defend. Alternating phases over three months on a realistic multi-agent system.