Soham Deshmukh

audio and language @Sesame Prev: @Microsoft @CarnegieMellon

[Sesame @sesame]

A collection of personal agents, crafted for everyday conversation.

Preview available now on iOS.

https://t.co/Hq5ZXAqPvX

[Satvik Dixit @SatvikDixit9]

Our paper "Mellow: a small audio language model for reasoning" was accepted to #NeurIPS2025!

Catch our poster in Session 3 at 11am today.

[Soham Deshmukh @sohamdesh_]

@HirofumiInaguma @rdesh26 Feel free to DM if you are looking for speech/audio roles

[Soham Deshmukh @sohamdesh_]

@MimansaJ Sent a DM

[Soham Deshmukh @sohamdesh_]

@animesh_garg @dwarkesh_sp I completely agree with the post. But I think the point here was to troll Lex 😂

[arXiv Sound @ArxivSound]

Shikhar Bharadwaj, Samuele Cornell, Kwanghee Choi, Satoru Fukayama, Hye-jin Shim, Soham Deshmukh, Shinji Watanabe, "OpenBEATs: A Fully Open-Source General-Purpose Audio Encoder," https://t.co/AsYCAjyw9q

[Dimitris Papailiopoulos @DimitrisPapail]

Some of the most impactful work you can do in academia isn’t cool new algos or novel architectures. It’s data research.

Data research isn’t just dumping tokens into a json. It requires a ton of rigorous experimentation, algorithmic thinking, and actually talking to your models.

[Soham Deshmukh @sohamdesh_]

@anishmaybe @AllTrails More importantly how was the hike?

[Soham Deshmukh @sohamdesh_]

Being presented currently in Hall 3 - poster number 48 - 10 am to 12:30 pm. Come drop by to discuss robustness evaluation for ASR systems! #ICLR2025

[Ahmed Shah @AhmedSh1494]

Happy to announce that Speech Robust Bench (SRB) is being presented at #ICLR2025. SRB is a comprehensive multi-lingual robustness benchmark for speech recognition.
paper: https://t.co/IDGOufnt1s
code: https://t.co/GPgfCIHZSD
data: https://t.co/EEbymoTNin
more in 🧵

[Soham Deshmukh @sohamdesh_]

@makemytripcare Absolutely worst customer service! My ticket was confirmed after boarding closed, forcing me to buy another ticket at the counter. After a week of chasing 5 different people, still no updates.

[Soham Deshmukh @sohamdesh_]

Just landed in Hyderabad for ICASSP 2025! Excited to reconnect with familiar faces and meet some new ones. Feel free to reach out or swing by the Microsoft booth for a quick chat #ICASSP2025

[Soham Deshmukh @sohamdesh_]

@harshit_sikchi @scottniekum @yayitsamyzhang @marcgbellemare @yukez @PeterStone_TX Congratulations Harshit! 🎉🎊

[Soham Deshmukh @sohamdesh_]

collaboration with @SatvikDixit9 and advised by rita singh and bhiksha raj

demo: https://t.co/65EWNVbdsQ
huggingface: https://t.co/hEEyLYfsFk
github: https://t.co/1ylRTSGG10
paper: https://t.co/ZOk7CX1xZ5

[Soham Deshmukh @sohamdesh_]

we show for the first time ever that sub-billion audio models can reason. we introduce mellow, a small audio-language model (167M) that gets SoTA on different audio reasoning tasks. by using our method and data, you can train an alm within 24 hrs on academic resources (1/n 🧵)

[arXiv Sound @ArxivSound]

``Mellow: a small audio language model for reasoning,'' Soham Deshmukh, Satvik Dixit, Rita Singh, Bhiksha Raj, https://t.co/2frf6F1gWq

[Soham Deshmukh @sohamdesh_]

also, if one scales data (audio), mellow performance improves further, highlighting the effectiveness of our minimal recipe. we release the checkpoint, reasonaqa dataset, and will soon open-source the training code to make research on audio-language models more accessible!

[𝚐𝔪𝟾𝚡𝚡𝟾 @gm8xx8]

Mellow: a small audio language model for reasoning

[arXiv Sound @ArxivSound]

``Mellow: a small audio language model for reasoning,'' Soham Deshmukh, Satvik Dixit, Rita Singh, Bhiksha Raj, https://t.co/2frf6F1gWq