🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Join us at the Oral session at #KDD2021 on Neural Instant Search for Music and Podcast. Follow the zoom link to attend the talk. virtual.2021.kdd.org/paper_session_…

NEW: This comprehensive report investigates foundation models (e.g. BERT, GPT-3), which are engendering a paradigm shift in AI. 100+ scholars across 10 departments at Stanford scrutinize their capabilities, applications, and societal consequences. bit.ly/3xZPFYK

Check our blog post about a recently presented paper at #KDD2021 on "Neural Instant Search for Music and Podcast" from Helia Hashemi (Spotify Research intern), Aasish Pappu (ఆశీష్ పప్పు), Mi Tian, Praveen Chandar, Mounia Lalmas & Ben Carterette. research.atspotify.com/neural-instant…

The "Vanishing Gradients" is a new Netflix series.

Do you want to work on - search relevance - recommendation - personalization using - machine-learning - embeddings - vector or hybrid retrieval but are kinda tired of all the plumbing? Thread time 👇

A blog post of recent work presented at #ICWSM2021 on the representation of music Creators on Wikipedia from colleagues Alice Wang Aasish Pappu (ఆశీష్ పప్పు) and No longer here. research.atspotify.com/representation…

Check the write-up of our #CIKM2021 paper "Leveraging Semantic Information to Facilitate the Discovery of Underserved Podcasts" from Maryam Aziz, Alice Wang, Aasish Pappu (ఆశీష్ పప్పు), Hugues Bouchard, @musicbug Ben Carterette Mounia Lalmas Spotify Research Spotify Engineering research.atspotify.com/researching-ho…

A bunch of people have been asking me about Spotify moving away from Luigi, the open source workflow scheduler Elias Freider and I built in 2011-2015:  engineering.atspotify.com/2022/03/why-we… – will put some thoughts in a thread

We are hiring research interns, postdocs, research scientists and engineers in multiple areas: on-device models, content understanding and generation, question answering, and summarization. Ou work span across different modalities focusing on AR and VR applications.

Today Meta AI is sharing OPT-175B, the first 175-billion-parameter language model to be made available to the broader AI research community. OPT-175B can generate creative text on a vast range of topics. Learn more & request access: ai.facebook.com/blog/democrati…

OPT-175B - making AI more open and reproducible, by @MetaAI. ai.facebook.com/blog/democrati…

I'm hiring 3 pre/postdoc researchers at EECS, UC Berkeley. Focus 1) assistive communication technologies fusing multiple biosensors. 2) Computational models of spoken language integrating brain, behavior and other biosignals. 3) Aphasia characterization from speech. Please RT!!

We are entering a new phase in generative models. Text-to-video is here! Make-a-video by @MetaAI and FAIR. makeavideo.studio Look at this video! It's generated! "A golden retriever eating ice cream on a beautiful tropical beach at sunset, high resolution"

Exciting opportunity! metacareers.com/jobs/642602210…

We are so sad that our dear friend Sreekanth Jaldi (IIITA B2K3) passed away on 11th December. His wife and two daughters (1 and 7 years old) are now facing a very hard future without their father. We come to you to ask you to help support Sreekanth’s family so his dreams can come

Nice article at The Verge about the AI-powered products announced by Meta today: chatbots with personalities, image and sticker generation, and more. theverge.com/2023/9/27/2389…

Here's my conversation with Mark Zuckerberg, his 3rd time on the podcast, but this time we talked in the Metaverse as photorealistic avatars. This was one of the most incredible experiences of my life. It really felt like we were talking in-person, but we were miles apart 🤯 It's

New Year Resolution for the #AI Prognosticator World: Every talking head pontificating about #AI dangers and impacts should be required to complete some kind of an Intro #AI course--so they don't let their imaginations run wilder than those of ChatGPT..🙃

🧵 Adapting your LLM for new tasks is dangerous! A bad training set degrades models by encouraging hallucinations and other misbehavior. Our paper remedies this for RAG training by replacing gold responses with self-generated demonstrations. Check it out: arxiv.org/abs/2502.10

Check out the full thread on bsky: bsky.app/profile/mattf1… Shout out to my collaborators at Meta Ilia Kulikov, Dan Bikel, Barlas Oğuz, Xilun Chen, Aasish Pappu (ఆశీష్ పప్పు), and a big thanks to the authors of the open-source library fairseq2 which made SFT and DPO training a piece of cake!