Meetween was at #Interspeech2026 in Sydney with three papers: a spoken prompt dataset for testing speech-language models, a method that cuts code-switching ASR errors by almost 33%, and new metrics for simultaneous speech translation latency. #SpeechAI #HorizonEurope

👩‍🔬 𝐓𝐡𝐞 𝐒𝐞𝐩𝐭𝐞𝐦𝐛𝐞𝐫 𝐞𝐝𝐢𝐭𝐢𝐨𝐧 𝐨𝐟 𝐭𝐡𝐞 𝐌𝐞𝐞𝐭𝐰𝐞𝐞𝐧 𝐌𝐨𝐧𝐭𝐡𝐥𝐲 𝐑𝐞𝐯𝐢𝐞𝐰 𝐢𝐬 𝐨𝐮𝐭. Researchers made 8 picks, covering real-time speech translation, voices in 600+ languages and how LLMs think: pischool.link/dme #Meetween #SpeechTechnology

📷 𝐌𝐞𝐞𝐭 𝐭𝐡𝐞 𝐄𝐱𝐩𝐞𝐫𝐭 𝐄𝐩𝐢𝐬𝐨𝐝𝐞 𝟑 Sebastian Stüker, Director of Research Science at @Zoom Germany, on founding a speech translation startup later acquired by Zoom, and how his work helps Meetween improve video meetings. #Meetween #SpeechTranslation #MeetTheExpert

Meetween Agentar: when you can't be there, your Agentar can. A virtual proxy that presents, discusses and represents you in meetings, bridging time zones and distance. Keep following Meetween! #Meetween #ai

📄 𝐌𝐞𝐞𝐭𝐰𝐞𝐞𝐧 𝐌𝐨𝐧𝐭𝐡𝐥𝐲 𝐑𝐞𝐯𝐢𝐞𝐰 𝐉𝐮𝐥𝐲 𝐢𝐬 𝐨𝐮𝐭. Three papers on long-context attention, full-duplex dialogue and speech-language pretraining, selected by TAUS, @ksri_kit and @fbk_mt. What connects them? Read here: pischool.link/tos #SpeechTechnology

🎙 𝐌𝐞𝐞𝐭 𝐭𝐡𝐞 𝐄𝐱𝐩𝐞𝐫𝐭 | 𝐄𝐩. 𝟐 Meet @Seb_Bratieres, Director of AI at @Translation & Meetween coordinator. He shares how his bilingual roots led him to language tech, and how Meetween brings speech translation closer to text quality. Watch below 👇 #Meetween #AI

Meetween's week in San Diego. KIT and FBK presented at #ACL2026 and #IWSLT2026 on full-duplex dialogue, #SpeechTranslation, audio chaptering and instruction following. Thanks to A. Waibel, F., Retkowski, J., Niehues, L., Bentivogli and M. Züfle. #Meetween

𝗠𝗲𝗲𝘁𝘄𝗲𝗲𝗻 𝗮𝘁 𝗔𝗖𝗟 𝟮𝟬𝟮𝟲 Machine Translation at FBK heads to ACL2026 in San Diego, with a paper presented at IWSLT2026. Findings on the Model Compression Task are also out now. 🔗 bit.ly/4gf7GMH 🔗 bit.ly/4wIHGOT #ACL2026 #IWSLT2026

📄 Meetween Monthly Review, June 2026 is out. Three papers on post-training strategies, cross-modality fusion, and low-resource speech recognition, selected by FBK and TAUS researchers. Read the full review 👇 pischool.link/MeetweenMR #Meetween #SpeechTechnology #Multimodal

🎙️ Meet the Expert | Ep.1 We met with Amir Kamran, a Solution Architect at TAUS, to discuss what attracted him to the field of language technology and TAUS's role as a data partner in @Meetween. Watch below 👇 #MeetTheExpert #MachineTranslation