Now that we have a frontier model that's open-weight (not open-source), it's time to go back to all those ambitious use cases where open-weight models failed to deliver (agents) and try again, so we can have reproducible science and not worry about API models getting deprecated.
And the Best Paper Award, sponsored by @juji_io goes to…Ashley Ge Zhang, Xiaohang Tang, Steve Oney, and Yan Chen for “CFlow: Supporting Semantic Flow Analysis of Students' Code in Programming Problems at Scale.” https://t.co/Fdt3E2PRbA
Congrats to all nominees! #las24ed 🎉🥳👏🏻
#3 At-Your-Own-Risk AI Agents Respond with General #LLMs 🎲
SMEs can't anticipate all user questions, & users may ask "out-of-the-box" queries 🤖
General LLMs handle unanswered questions but risk AI hallucinations 🔄 @juji_io allows toggling on/off LLM responses for flexibility
Control and trace AI hallucinations in #generativeAI with Juji's 3 AI Chatbot Safety Scales for Organizations! 🚨 Ensure accurate information in education & healthcare with zero tolerance for errors. 🔍 Leverage the power of #genAI safely! 🧠
#2 Semi-Safe AI Agents Respond with #RAG 🧩
🔍 Retrieval-Augmented Generation augments #LLMs by generating responses from proprietary data ⚠️ RAG inherits hallucinations from general LLMs 📈
Juji offers options to disable RAG or monitor & audit for improvements
#1 Safest AI Agents Respond with Human-Curated Knowledge 🧠
💡 SMEs directly author answers or audit AI-generated responses 🔒 AI agents use only curated knowledge to ensure accuracy
Experience the safest #AI with @juji_io ! 🚀
We should call models like Llama 3, Mixtral, etc. “open-weight models”, not “open-source models”. For a model to be open-source, the code and training data need to be public (good examples: GPT-J, OLMo, RedPajama, StarCoder, K2, etc.). Weights are like an exe file, which would be ridiculous to call open-source.
When a proactive AI agent asks a series of questions, it should do so in a conversational tone with smooth transitions, like an articulate human interviewer. Juji AI Studio lets designers auto-generate and edit such flows using generative AI for natural conversations. #Chatbots
Furthermore, for each expression, a conversational AI designer can indicate whether it is only used for re-asking, giving you fine-grained control to make an AI agent articulate.
A proactive AI agent doesn't just passively respond to users. It actively guides them toward an agenda, like tutoring or coaching, while being articulate in asking and answering questions. Just like a good human agent! 🤖💬
#AI#Chatbots#ProactiveAI#Coaching#Tutoring
A conversational AI designer can use Juji Studio to customize an AI agent’s message, including paraphrasing a question either manually or automatically using generative AI.
Dear all,
I need a few emergency reviewers for manuscripts related to (1) handwriting recognition; (2) common NLP tasks (e.g. QA, NER). If you have strong publications in these areas and have bandwidth to review within the next 2-3 weeks, please DM me. Thanks.
Andrew Ng discusses "AI agentic workflows" where agents and humans engage in multi-turn interactions to achieve a common goal. Juji enables this to automate high-touch engagements like advising, career dev, and healthcare counseling. 🤝🤖
https://t.co/GLulwuMDhO
#AI#aiagents
Juji's AI powered tutor won Distinguished Innovation Award from International Society of Service Innovation Professionals @The_ISSIP
https://t.co/BhQS9bigTo
Guiding students through @FAFSA with an AI chatbot is complex due to potential side conversations at any point. Juji allows designers to focus on the main workflow while automatically managing side conversations, no matter how complex. No CS degree needed!🎓🤖#AI#Chatbots#FAFSA
Filling out the FAFSA for college financial aid can be a nightmare. An AI advisor could guide students through the process based on their specific situation. Who should partner to help students navigate this maze and fund their education? 🎓💰
#FAFSA#collegelife#AI