Cohere For AI

Tuesday, October 8th, join our Reinforcement Learning group as they welcome Panayiotis Panayiotou for a conversation on "Curricula for Learning Robust Policies with Factored State Representations in Changing Environments!" 🤔 Thank you Rahul Narava Gusti Triandi Winata and Djamel Rassem Lamouri for organizing this event! 🤗 Learn more: https://lnkd.in/gAdgUcbd

The October C4AI Newsletter is here! 🚀 Building off the success of Aya, we recently wrapped up Expedition Aya: a six-week, global open-build challenge to connect researchers worldwide and support new multilingual research initiatives.  Join us in celebrating the five projects that were honored with the top recognition: https://lnkd.in/gXN_HSJd 🎉C4AI celebrates over 100 guest speakers with our open science community. Thank you to all the community leads who dedicate their time and efforts to organize these events. Check out previous sessions here! https://lnkd.in/gndNtnTm 🌟C4AI Research Scientist, Marzieh Fadaee, recently gave the closing keynote for PyData Amsterdam. Marzieh spoke on The Art of Language: Mastering Multilingual Challenges in LLMs. “It was fantastic to discuss how we built state-of-the-art multilingual models and datasets through our open science Aya initiative, driving innovation and collaboration in the research community.” - Marzieh Fadaee 🛣️Roads to Research - This month join us tomorrow, as Bryan M. Li sits down with Winnie X., Gbemileke Onilude and Emma Strubell to talk us through the Paths to Graduate Studies. 📚  Learn more here:https://lnkd.in/gPd-x8Ba And October 29, join us for a discussion on Mentorship in ML research with Pablo Samuel Castro, Esra'a Saleh and Viraat Aryabumi! Learn more here:https://lnkd.in/g-guJ39g 👏Our team is excited to connect, share ideas and learn with you at EMNLP and NeurIPS 2024! We want to congratulate all those who have had their work accepted to the conferences. EMNLP Acceptances 🔸RLHF Can Speak Many Languages: Unlocking Multilingual Preference Optimization for LLMs https://lnkd.in/g5q4hHgr 🔸LLM See, LLM Do: Guiding Data Generation to Target Non-Differentiable Objectives https://lnkd.in/gU5V6MUV 🔸The Multilingual Alignment Prism: Aligning Global and Local Preferences to Reduce Harm https://lnkd.in/gsNQnTkS NeurIPS Acceptances 🔸Elo Uncovered: Robustness and Best Practices in Language Model Evaluation https://lnkd.in/g2swjTYJ 🔸BAM! Just Like That: Simple and Efficient Parameter Upcycling for Mixture of Experts https://lnkd.in/gXfpAnJ9 ✨This month we are spotlighting @Ameed Taylor, one of the co-leads for our Open Science Community’s Computer Vision group. Ameed is a Cloud ERP & Analytics architect and Generative AI Specialist! 📩 Sign up to get these monthly newsletters straight to your inbox: https://lnkd.in/gsFGDF44 

Tomorrow, you do not want to miss this Roads to Research session with Bryan M. Li who will be hosting a panel about paths to graduate studies with Emma Strubell, Winnie X. and Gbemileke Onilude!

Tomorrow, we're looking forward to this session with our NLP group and Cong Lu! 👨🔬

One of the standout moments for me at PyData Amsterdam 2024 was the inspiring keynote by Marzieh Fadaee, a senior research scientist at Cohere For AI. She introduced us to Project Aya, an ambitious initiative that's set to transform how we interact with AI across the globe. Why is Project Aya a Game-Changer? 🌎 Global Collaboration: Over 3,000 independent researchers from 119 countries have united to bring Aya to life. This massive collaborative effort exemplifies the power of a global community working towards a common goal. 🗣️ Supporting 100+ Languages: Aya embraces over 100 languages, including many underrepresented in AI. This inclusivity makes AI technology accessible to a broader spectrum of cultures and communities. 🔓 Fully Open-Sourced: With both the dataset and model openly available, Aya fosters innovation by allowing worldwide collaboration without barriers. Why Does This Matter? By embracing such a wide array of languages, Project Aya is not just a technological achievement—it's a cultural bridge. Here's what this means: 💪 Empowering Underrepresented Communities: Aya provides tools that enable communities to access information and technology in their native tongues. 🌿 Preserving Cultural Heritage: Languages carry the essence of cultures. Aya's support aids in preserving linguistic diversity for future generations. 🤖 Enhancing AI Interactions: A multilingual AI can understand and respond to users more naturally, improving user experience across different linguistic backgrounds. Real-World Impact Across Industries 🌐 Healthcare: Aya enhances patient communication by powering multilingual AI assistants in medical consultations. 💼 Recruiting and HR: Companies can tap into diverse talent pools using Aya's ability to process multiple languages. 📚 Education: Learning platforms can offer materials in various languages, making education more accessible globally. Impressive Performance 📊 Aya has demonstrated competitive results in multilingual benchmarks, outperforming many popular LLMs, especially in underrepresented languages. This showcases its capability to effectively understand and process a diverse range of languages. 🚀 Get Involved and Explore If you're intrigued by Project Aya and its mission, here are some ways to dive deeper: 🔗 Discover Project Aya: cohere.com/research/aya 🔗 Try it out: https://lnkd.in/ePtfFpwx #PyData2024 #ProjectAya #MultilingualAI #DataScience #AIforGood #Inclusivity #Diversity #OpenSource #NLP #Collaboration #AIforAll #LanguageDiversity

Looking to explore new methods in multi-image visual question answering? 🖼️ Join our community-led Computer Vision group with Tsung-Han (Patrick) Wu for a presentation October 8th! Thank you Ameed Taylor and Benedict Emoe-kabu for organizing this event! 👏

🤖 Please join the Cohere For AI Computer Vision community for guest speaker Tsung-Han (Patrick) Wu on 𝐓𝐮𝐞𝐬𝐝𝐚𝐲 𝐎𝐜𝐭𝐨𝐛𝐞𝐫 𝟖𝐭𝐡 𝐚𝐭 𝟏𝟏:𝟎𝟎 𝐚𝐦 𝐄𝐃𝐓.🤖 Patrick is a EECS PhD student at UC Berkeley, advised by Prof. Trevor Darrell, affiliated with BAIR (https://bair.berkeley.edu) and Prof. Joseph E. Gonzalez in Sky computing Lab (https://lnkd.in/d8E2VBse In this session, Patrick will introduce new methods in Multi-Image Visual Question Answering, focusing on retrieval and reasoning across diverse image sets using the Visual Haystacks benchmark.📘 🗓️ 𝐒𝐞𝐬𝐬𝐢𝐨𝐧 𝐋𝐢𝐧𝐤: https://lnkd.in/dx2M43H2 📖 𝐏𝐚𝐩𝐞𝐫 𝐋𝐢𝐧𝐤: https://lnkd.in/dpsM2T45 👁️ 𝐏𝐚𝐭𝐫𝐢𝐜𝐤'𝐬 𝐖𝐞𝐛 𝐒𝐢𝐭𝐞: https://lnkd.in/d2y7ZnuK. #computervision #cohereforai #UCberkeley #VisualAI #AIResearch #VisionModels

想知道 🧭 🧨 📜 🖨️ 有什么共同之处吗？来问Aya吧！ Aya 23 is available in Chinese! ✨ Try it now: https://lnkd.in/gp4ZxWJv

Tomorrow join our Geo Regional Asia group with Ziqi Wang and hear about teaching LMs self-correction ✅ !

This Saturday tune into the next 🛣️ Roads to Research session with Bryan M. Li and learn about the different perspectives on paths to graduate studies!