Ziyang Ma@ICML26🇰🇷

PhD @sjtu1896 @NTUsg | Speech, Language, Audio, Music AI | Now @TencentHunyuan Multimodal | Ex @Alibaba_Qwen Qwen-Omni @BytedanceTalk Seed @MSFTResearch MSRA

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

If you or your team are looking for someone with my background, or if you just want to grab a coffee and chat about multimodal AI in Seoul, please feel free to reach out via DM or connect directly!
See you! 🇰🇷

#ICML2026 #MultimodalAI #SpeechAI #OmniModels #FullDuplex #Hiring

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

🚀 I have one paper accepted, and I will be attending ICML 2026 in Seoul!

As an AI researcher deeply passionate about the next generation of human-AI interfaces, my research and professional interests lie at the intersection of:

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

- Speech & Audio-Video Omni-Models
- Proactive Interaction Models & Full-Duplex Modeling
- Unified Multimodal Foundation Models

💡 I am currently actively seeking full-time positions. I would love to connect with fellow researchers, industry professionals, and recruiters.

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

@_akhaliq

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

Let us cook a speech & audio Banana🍌

[Tencent Hy @TencentHunyuan]

Can AI truly edit audio, not just generate it? 🎧

Tencent Hy, in collaboration with SJTU, SII, NTU, TJU, ZODA, PKU, FDU, and other collaborators, introduces MMAE.

MMAE--A Massive Multitask Audio Editing Benchmark, is the first comprehensive evaluation benchmark for speech and audio "Banana🍌"

Instead of simply requiring the AI to "generate" audio, it demands that the AI understand an existing audio clip and precisely modify it according to natural language instructions—altering what needs to be changed while leaving the rest untouched.

Current models show an Exact Match Rate (EMR) below 5%, revealing a major gap in reliable audio editing.

MMAE includes:
✅ 2,000 high-fidelity samples from real-world scenarios
✅ 17,741 fine-grained rubric evaluation items
✅ 7 modality settings across sound, music, speech and their mixtures
✅ 6 task complexity from basic modifications to multi-hop reasoning and multi-round editing
✅ 8 operation types across local and global granularities

How to use:
arXiv: https://t.co/TM81ahH7PZ
GitHub: https://t.co/UR1dRUKqMD
HuggingFace: https://t.co/1MHR1n3LJn
Demo: https://t.co/tz2TVHaCk8

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

Welcome to our session on pushing speech & audio, time-series, and EEG reasoning. I won't be at ICASSP in person this year, but you can reach out to Yohei Kawaguchi @yohekawag , Xie Chen, and other organiers to connect!

[Yohei Kawaguchi @yohekawag]

Heading to #ICASSP2026 !
Excited to announce a special session we’re organizing:
SS-L10: Question Answering and Reasoning on Audio and Time-Series Data
📅 Thu, 7 May | 14:00–16:00
📍 Room 111
🔗 https://t.co/kkd5qa4IpD
Please join us and let’s have a great discussion! #ICASSP

[Qwen @Alibaba_Qwen]

Demo1：Audio-Visual Captioning

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

Check it out!

[Tongyi Lab @Ali_TongyiLab]

1/10 🚀 Qwen3.5-Omni is here! Scaling up to a native omni-modal AGI.
Meet the next generation of Qwen, designed for native text, image, audio, and video understanding, with major advances in both intelligence and real-time interaction.
A standout feature:
Audio-Visual Vibe Coding: Describe your vision to the camera, and Qwen3.5-Omni instantly builds a functional website or game for you.
Highlights:
Script-Level Captioning: Generate detailed video scripts with timestamps, scene cuts & speaker mapping.
SOTA Performance: Qwen3.5-Omni has secured 215 SOTA scores across various sub-tasks, matching the top-tier text/vision capabilities of the Qwen3.5 series.
Audio-Visual Understanding: From auto-segmentation to fine-grained script generation, it understands the relationship between characters and their environment like never before.
Seamless Interaction: With native API support for Semantic Interruption, voice conversations feel human-like and background-noise resistant.
Global Multilingual Mastery: Pioneering support for 74 languages in speech recognition and 29 languages in expressive speech generation, breaking down global communication barriers.
Autonomous Intelligence: Native support for WebSearch and complex Function Calling—the model now independently decides when to pull real-time data.
Qwen3.5-Omni is built to be the backbone of next-gen AI applications, empowering developers and users alike with true multimodal reasoning.

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

[5/5] Resources:
🔗 Code: https://t.co/Kby6zNa1To
🔗 Data: https://t.co/LkLXX8XJIn
🔗 Paper: https://t.co/qSJBJCU9GC
🔗 Demo: https://t.co/F2xhldh7Jl

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

[1/5] Excited to announce the open-source release of Omni-Cloze! 🚀
A fine-grained audio-visual captioning benchmark designed to solve the "hallucination vs. detail" trade-off in Omni-modal models.

#AI #Multimodal #LLM #Qwen #OmniCaptioner #OmniCloze #ICLR2026 #OpenSource

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

[4/5]
3️⃣ Massive Scale: 2k clips and 70k fine-grained questions across 47 categories (Tech, Sports, Education, etc.).
4️⃣ Human-Level Accuracy: Hits a 0.91 Pearson correlation with human judgment (Elo)—surpassing VDC and video-SALMONN 2!

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

@JustinLin610 keep changing the world!

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

We are excited to announce the leaderboard is officially live! The organizing committee is also thrilled to announce a special prize: an Nvidia RTX 5070 GPU! This will be awarded to the "Best Innovative Award" winner, recognizing the most creative single-model or agent system.

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

We are thrilled to announce that the Audio Reasoning Challenge will be officially held at Interspeech 2026 !
Link to challenge website: https://t.co/ZuSRny7f2B

[Qwen @Alibaba_Qwen]

🚀 Meet the new Qwen3-TTS lineup: VoiceDesign & VoiceClone!
Create, control, and clone voices—faster and more expressive than ever.
⚡ VoiceDesign-VD-Flash
• Fully controllable speech via free-form text instructions — tone, rhythm, emotion, persona
• No preset voices. Design your own unique vocal identity
• Outperforms GPT-4o-mini-tts & Gemini-2.5-pro on role-play benchmarks
⚡ VoiceClone-VC-Flash
• Clone any voice from just 3 seconds of audio
• Generate speech in 10 languages (CN / EN / JP / ES + more)
• 15% lower WER vs. ElevenLabs & GPT-4o-Audio in multilingual tests
• Context-aware cadence for more natural delivery
🎙️ Try it now
• Qwen Chat: https://t.co/nnAW9ZfjoV
• Blog: https://t.co/8yYEDPf2iC

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

Just finished my amazing trip for #NeruIPS2025 at San Diego and heading to #ASRU2025 at Hiwikii!

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

Yinghao and I will be at NeurIPS and Jinyu and I will be at ASRU this month. Drop by if you are interested in it!

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

We are thrilled to announce that the Audio Reasoning Challenge will be officially held at Interspeech 2026 !
Link to challenge website: https://t.co/ZuSRny7f2B

[Ziyang Ma@ICML26🇰🇷 @ddlbojack]

@holynski_ Hi, do you have positions related to speech/audio/omni models? thx