In recent years, large language models (LLMs) have shown an impressive ability to perform arithmetic and symbolic reasoning tasks. However, we found that LLMs (e.g., ChatGPT) cannot perform well on reasoning that requires multiple rounds of dialogue, especially when solving situation puzzles. Specifically, LLMs intend to ask very detailed questions focusing on a specific aspect or same/similar questions after several rounds of Q&As. To help LLMs get out of the above dilemma, we propose a novel external reformulation methodology, where the situation puzzle will be reformulated after several rounds of Q&A or when the LLMs raise an incorrect guess. Experiments show superior performance (e.g., win rate, number of question/guess attempts) of our method than directly using LLMs for solving situation puzzles, highlighting the potential of strategic problem reformulation to enhance the reasoning capabilities of LLMs in complex interactive scenarios.

Text-to-SQL systems have unlocked easier access to critical data insights by enabling natural language queries over structured databases. However, deploying such systems in enterprise environments remains challenging due to factors such as large, complex schemas (> 3000 columns), diverse SQL dialects (e.g., BigQuery, Snowflake) and sophisticated query requirements (e.g., transformation, analytics). Current state-of-the-art performance on the Spider 2.0 dataset -- a benchmark built to mimic such complex environments -- remains limited at 20%. Key limitations include inadequate instruction-following, poor long-context comprehension, weak self-refinement, and insufficient dialect-specific knowledge. To address these gaps, we propose ReFoRCE (Self-Refinement Agent with Format Restriction and Column Exploration) which introduces (1) table compression to mitigate longcontext limitations (2) format restriction to ensure accurate answer format, and (3) iterative column exploration for enhanced schema understanding. Additionally, it employs self-refinement pipeline consisting of (1) parallelized workflows with voting mechanisms and (2) a Common Table Expression (CTE) based refinement approach to handle unresolved cases. ReFoRCE achieves state-of-the-art results scoring 31.26 on the Spider 2.0-Snow and scoring 30.35 on the Spider 2.0-Lite tasks. Text-to-SQL converts natural language queries into SQL queries, serving as a key technology for lowering the barrier to accessing relational databases (Zelle & Mooney, 1996; Zettlemoyer & Collins, 2012; Zhong et al., 2017; Yu et al., 2018; Wang et al., 2019; Gao et al., 2023a; Lei et al., 2024).

The 2023 AI4EIC hackathon was the culmination of the third annual AI4EIC workshop at The Catholic University of America. This workshop brought together researchers from physics, data science and computer science to discuss the latest developments in Artificial Intelligence (AI) and Machine Learning (ML) for the Electron Ion Collider (EIC), including applications for detectors, accelerators, and experimental control. The hackathon, held on the final day of the workshop, involved using a chatbot powered by a Large Language Model, ChatGPT-3.5, to train a binary classifier neutrons and photons in simulated data from the \textsc{GlueX} Barrel Calorimeter. In total, six teams of up to four participants from all over the world took part in this intense educational and research event. This article highlights the hackathon challenge, the resources and methodology used, and the results and insights gained from analyzing physics data using the most cutting-edge tools in AI/ML.

With the advent of Large Language Models (LLM), conversational assistants have become prevalent for domain use cases. LLMs acquire the ability to contextual question answering through training, and Retrieval Augmented Generation (RAG) further enables the bot to answer domain-specific questions. This paper describes a RAG-based approach for building a chatbot that answers user's queries using Frequently Asked Questions (FAQ) data. We train an in-house retrieval embedding model using infoNCE loss, and experimental results demonstrate that the in-house model works significantly better than the well-known general-purpose public embedding model, both in terms of retrieval accuracy and Out-of-Domain (OOD) query detection. As an LLM, we use an open API-based paid ChatGPT model. We noticed that a previously retrieved-context could be used to generate an answer for specific patterns/sequences of queries (e.g., follow-up queries). Hence, there is a scope to optimize the number of LLM tokens and cost. Assuming a fixed retrieval model and an LLM, we optimize the number of LLM tokens using Reinforcement Learning (RL). Specifically, we propose a policy-based model external to the RAG, which interacts with the RAG pipeline through policy actions and updates the policy to optimize the cost. The policy model can perform two actions: to fetch FAQ context or skip retrieval. We use the open API-based GPT-4 as the reward model. We then train a policy model using policy gradient on multiple training chat sessions. As a policy model, we experimented with a public gpt-2 model and an in-house BERT model. With the proposed RL-based optimization combined with similarity threshold, we are able to achieve significant cost savings while getting a slightly improved accuracy. Though we demonstrate results for the FAQ chatbot, the proposed RL approach is generic and can be experimented with any existing RAG pipeline.

Castle Romeo was the code name given to one of the tests in the Operation Castle series of American ... [ ] thermonuclear tests beginning in March 1954 at Bikini Atoll. The ultimate objective was to test designs for an aircraft deliverable thermonuclear weapon. Microsoft announced it was placing new limits on its Bing chatbot following a week of users reporting some extremely disturbing conversations with the new AI tool. The chatbot expressed a desire to steal nuclear access codes and told one reporter it loved him. "Starting today, the chat experience will be capped at 50 chat turns per day and 5 chat turns per session. A turn is a conversation exchange which contains both a user question and a reply from Bing," the company said in a blog post on Friday.

Microsoft has limited the number of "chat turns" you can carry out with Bing's AI chatbot to five per session and 50 per day overall. Each chat turn is a conversation exchange comprised of your question and Bing's response, and you'll be told that the chatbot has hit its limit and will be prompted to start a new topic after five rounds. The company said in its announcement that it's capping Bing's chat experience because lengthy chat sessions tend to "confuse the underlying chat model in the new Bing." Indeed, people have been reporting odd, even disturbing behavior by the chatbot since it became available. New York Times columnist Kevin Roose posted the full transcript of his conversation with the bot, wherein it reportedly said that it wanted to hack into computers and spread propaganda and misinformation.

Microsoft is planning to update Bing Chat to make it a little less bizarre. OpenAI, meanwhile, is trying to make ChatGPT a bit more neutral. Microsoft issued a blog post late Wednesday outlining the changes coming to Bing's AI chatbot, while OpenAI issued a technical blog post on Thursday with more details on how it was tuning its model, too. It's hard to believe that the new Bing Chat has only been out a week, but the new Bing has gained a popularity that it has rarely, if ever had. In a blog post, Microsoft pointed to the "increased engagement" that Bing has seen as both its updated search and the Bing Chat AI chatbot have debuted in 169 countries.

The times have changed & customer experience (CX) is a major source of income and growth for businesses around the world. As the discipline has increased in relevance and legitimacy, so has the number of people working in it. At family dinners, CX is no longer relegated to the kids' table; it has taken its proper place at the head of the table. Customers have high expectations, and businesses understand the value of investing in CX. Here are some interesting statistics that will boost your excitement for CX.

In order to build a working full-stack application, there are so many moving parts to think about. And you'll need to make many decisions that will be critical to the success of your app. For example, what language will you use and what platform will you deploy on? Are you going to deploy a containerised software on a server, or make use of serverless functions to handle the backend? Do you plan to use third-party APIs to handle complex parts of your application, like authentication or payments? Where do you store the data? In addition to all this, you'll also need to think about the user interface, design and usability of your application, and much more. This is why complex large applications require a multifunctional development team collaborating to build the app. One of the best ways to learn how to develop full stack applications is to build projects that cover the end-to-end development process. You'll go through designing the architecture, developing the API services, developing the user interface, and finally deploying your application. So this tutorial will take you through the process of building an AI chatbot to help you learn these concepts in depth. Important Note: This is an intermediate full stack software development project that requires some basic Python and JavaScript knowledge. I've carefully divided the project into sections to ensure that you can easily select the phase that is important to you in case you do not wish to code the full application.

Prior work on sentiment analysis using weak supervision primarily focuses on different reviews such as movies (IMDB), restaurants (Yelp), products (Amazon).~One under-explored field in this regard is customer chat data for a customer-agent chat in customer support due to the lack of availability of free public data. Here, we perform sentiment analysis on customer chat using weak supervision on our in-house dataset. We fine-tune the pre-trained language model (LM) RoBERTa as a sentiment classifier using weak supervision. Our contribution is as follows:1) We show that by using weak sentiment classifiers along with domain-specific lexicon-based rules as Labeling Functions (LF), we can train a fairly accurate customer chat sentiment classifier using weak supervision. 2) We compare the performance of our custom-trained model with off-the-shelf google cloud NLP API for sentiment analysis. We show that by injecting domain-specific knowledge using LFs, even with weak supervision, we can train a model to handle some domain-specific use cases better than off-the-shelf google cloud NLP API. 3) We also present an analysis of how customer sentiment in a chat relates to problem resolution.