University of Michigan - Center for Academic Innovation的动态

We had a wonderful time celebrating the launch of Learning Languages with AI with faculty partner Pamela Bogart! ????Congrats Pamela Bogart, Hedieh Najafi, Ph.D. Ben Corona, Weiyi Zhang, Clare Brown, Junhee Chung, Lya Li, Briana Johnson, Molly Bodner, Sean Corp, Ben Anderson, Zhixin Li, Sam Baldwin ??Learning a new language can be exciting, but we don’t always have the resources to push ourselves to the next level. ??In “Learning Languages with AI”, you will learn how to enhance language learning with the help of generative AI tools. Identify your strengths and areas for improvement, then explore how you can responsibly harness AI to enhance your language skills even further. https://lnkd.in/exxZMyu8

?? Check out our recently submitted paper, "The global landscape of academic guidelines for generative AI and Large Language Models," co-authored with scholars from UT, IBM, and Allen AI2. We explore the benefits and ethical complexities of integrating Generative AI and LLMs in academia through a comprehensive survey and text-mining analysis of global and national directives and university guidelines. Discover how we can foster responsible innovation and ethical practices in education: https://lnkd.in/gHNCSh78 #AI #GenerativeAI #LLM #EthicsInAI #AcademicResearch

Check out the new paper, "ASTPrompter: Weakly Supervised Automated Language Model Red-Teaming to Identify Likely Toxic Prompts." This research introduces a reinforcement learning approach to red-teaming large language models, enabling the discovery of prompts that trigger toxic outputs while maintaining low perplexity. The paper provides a detailed analysis of the techniques used, making it a valuable resource for developers and researchers in AI and language modeling. The findings highlight significant improvements in identifying likely toxic prompts in a realistic setting. For those interested in the technical aspects and practical applications of language model robustness, this paper is essential reading. It offers insights and solutions that contribute meaningfully to advancements in the field. Read the full paper here: [https://lnkd.in/gC8PNBJc] #AI #MachineLearning #LanguageModeling #DataScience #Research #Innovation #RedTeaming #ToxicityDetection

?? Excited to share a new blog post on "Embedding-Aligned Language Models." This research introduces a revolutionary method for training large language models using reinforcement learning to steer text generation towards optimal regions of the latent embedding space. The EAGLE agent demonstrated its effectiveness in surfacing content gaps that align with latent user demand. Check out the full abstract here: https://bit.ly/3yO6oEM #language #AI #textgeneration

Future of Language Learning with LLMs https://lnkd.in/git9rbnP Discover how Large Language Models (LLMs) and AI are shaping the future of language education. Experience personalized language learning like never before! #LanguageLearningwithLLMs #AIinLanguageLearning #LanguageLearningTechnology #PersonalizedLanguageLearning #AI #AINews #AnalyticsInsight #AnalyticsInsightMagazine

Research is really boring and stressful. That's what I used to think. But that changed this semester. This semester Sanket Shivashankaran and I worked on an AI-based research project under the guidance of Kaustubh Kundu and Kalin Stefanov . Imagine being able to convert sign language --> text and vice versa in real-time! (That was the end goal of the project) Though, we worked on only one aspect of the project, which involved, using AI to improve sign language recognition. Our goal was to - ?? Test the accuracy of existing sign language recognition models and ?? Test whether additional models like deblurring and occlusion can be used to improve the accuracy of the sign language models. Installing large packages and the long testing times were part of the process and needed a lot of patience. Really grateful for the continuous guidance and support provided by Kaustubh Kundu throughout the project. I got to explore my curiosity in AI and learn a lot about it through the project. Overall, it wasn't easy tho but definitely interesting. I'm glad I could contribute to a project that can make industries more inclusive! #AI #ArtificialIntelligence #Technology #SignLanguage #research

Pleased to announce the release of the book "Talking Tech: The Engineering Behind Conversational AI Systems", which I had the opportunity to edit. ?? About the Book This book provides a detailed look into conversational AI, covering topics such as: Natural dialogue design The role of engineering students in AI innovation Applications of conversational AI across industries Insights into Large Language Models (LLMs) and implementation strategies ?? Details Publisher: Notion Press ISBN: 9798896739180 Grateful to Parul University and all who have contributed to this project and looking forward to its impact on students, educators, and professionals. #AI #ConversationalAI #BookRelease #ArtificialIntelligence

A must read if you want to get more about LLMs and language learning powered by AI.

If Artificial Intelligence continues to improve over the next few years, it will have a major impact on the way we learn languages. If vocabulary apps incorporate AI, they will be able to target our weaknesses. If teachers use AI, they will be able to plan different types of lessons. The likely consequences of AI on language learning are significant. Today I want to talk about Artificial Intelligence and its likely impact on language learning and language learners. To do this, we need to understand the?first conditional?and know how to use the first conditional in real life situations. https://lnkd.in/e9dbm6ex

A Recent Stanford Study Evaluates the Evolution of Multimodal Foundation Models from Few-Shot to Many-Shot-In-Context Learning In the dynamic landscape of large language models (LLMs) and large multimodal models (LMMs), the integration of demonstrating examples, known as in-context learning (ICL), has emerged as a pivotal enhancement strategy. Recent research from Stanford University sheds light on the transformative impact of incorporating in-context learning, particularly in the realm of multimodal models scaling from few-shot to many-shot scenarios. This investigation, led by a team of researchers, delves into the efficacy of advanced models such as GPT-4o and Gemini 1.5 Pro, shedding light on their performance dynamics in response to increasing demonstrating examples. https://is.gd/d6Jk9m #AI #AItechnology #artificialintelligence #incontextlearning #llm #LMMs #machinelearning