This paper introduces a novel approach to enhance Large Language Models (LLMs) with expert knowledge to automate the analysis of corporate sustainability reports by benchmarking them against the Task Force for Climate-Related Financial Disclosures (TCFD) recommendations. Corporate sustainability reports are crucial in assessing organizations' environmental and social risks and impacts. However, analyzing these reports' vast amounts of information makes human analysis often too costly. As a result, only a few entities worldwide have the resources to analyze these reports, which could lead to a lack of transparency. While AI-powered tools can automatically analyze the data, they are prone to inaccuracies as they lack domain-specific expertise. This paper introduces a novel approach to enhance LLMs with expert knowledge to automate the analysis of corporate sustainability reports. We christen our tool CHATREPORT, and apply it in a first use case to assess corporate climate risk disclosures following the TCFD recommendations. CHATREPORT results from collaborating with experts in climate science, finance, economic policy, and computer science, demonstrating how domain experts can be involved in developing AI tools. We make our prompt templates, generated data, and scores available to the public to encourage transparency.

With the remarkable generation performance of large language models, ethical and legal concerns about using them have been raised, such as plagiarism and copyright issues. For such concerns, several approaches to watermark and detect LLM-generated text have been proposed very recently. However, we discover that the previous methods fail to function appropriately with code generation tasks because of the syntactic and semantic characteristics of code. Based on \citet{Kirchenbauer2023watermark}, we propose a new watermarking method, Selective WatErmarking via Entropy Thresholding (SWEET), that promotes "green" tokens only at the position with high entropy of the token distribution during generation, thereby preserving the correctness of the generated code. The watermarked code is detected by the statistical test and Z-score based on the entropy information. Our experiments on HumanEval and MBPP show that SWEET significantly improves the Pareto Frontier between the code correctness and watermark detection performance. We also show that notable post-hoc detection methods (e.g. DetectGPT) fail to work well in this task. Finally, we show that setting a reasonable entropy threshold is not much of a challenge. Code is available at https://github.com/hongcheki/sweet-watermark.

Despite the many use cases for large language models (LLMs) in creating personalized chatbots, there has been limited research on evaluating the extent to which the behaviors of personalized LLMs accurately and consistently reflect specific personality traits. We consider studying the behavior of LLM-based agents, referred to as LLM personas, and present a case study with ChatGPT and GPT-4. The study investigates whether LLMs can generate content that aligns with their assigned personality profiles. To this end, we create distinct LLM personas based on the Big Five personality model, have them complete the 44-item Big Five Inventory (BFI) personality test and a story writing task, and then assess their essays with automatic and human evaluations. Results show that LLM personas' self-reported BFI scores are consistent with their designated personality types, with large effect sizes observed across five traits. Additionally, there are significant correlations between the assigned personality types and certain psycholinguistic features of their writings, as measured by the Linguistic Inquiry and Word Count (LIWC) tool. Interestingly, human evaluators perceive the stories as less personal when told that the stories are authored by AI. However, their judgments on other aspects of the writing such as readability, cohesiveness, redundancy, likeability, and believability remain largely unaffected. Notably, when evaluators were informed about the AI authorship, their accuracy in identifying the intended personality traits from the stories decreased by more than 10% for some traits. This research marks a significant step forward in understanding the capabilities of LLMs to express personality traits.

Scientific literature review generation aims to extract and organize important information from an abundant collection of reference papers and produces corresponding reviews while lacking a clear and logical hierarchy. We observe that a high-quality catalogue-guided generation process can effectively alleviate this problem. Therefore, we present an atomic and challenging task named Hierarchical Catalogue Generation for Literature Review as the first step for review generation, which aims to produce a hierarchical catalogue of a review paper given various references. We construct a novel English Hierarchical Catalogues of Literature Reviews Dataset with 7.6k literature review catalogues and 389k reference papers. To accurately assess the model performance, we design two evaluation metrics for informativeness and similarity to ground truth from semantics and structure.Our extensive analyses verify the high quality of our dataset and the effectiveness of our evaluation metrics. We further benchmark diverse experiments on state-of-the-art summarization models like BART and large language models like ChatGPT to evaluate their capabilities. We further discuss potential directions for this task to motivate future research.

Agents capable of carrying out general tasks on a computer can improve efficiency and productivity by automating repetitive tasks and assisting in complex problem-solving. Ideally, such agents should be able to solve new computer tasks presented to them through natural language commands. However, previous approaches to this problem require large amounts of expert demonstrations and task-specific reward functions, both of which are impractical for new tasks. In this work, we show that a pre-trained large language model (LLM) agent can execute computer tasks guided by natural language using a simple prompting scheme where the agent Recursively Criticizes and Improves its output (RCI). The RCI approach significantly outperforms existing LLM methods for automating computer tasks and surpasses supervised learning (SL) and reinforcement learning (RL) approaches on the MiniWoB++ benchmark. We compare multiple LLMs and find that RCI with the InstructGPT-3+RLHF LLM is state-of-the-art on MiniWoB++, using only a handful of demonstrations per task rather than tens of thousands, and without a task-specific reward function. Furthermore, we demonstrate RCI prompting's effectiveness in enhancing LLMs' reasoning abilities on a suite of natural language reasoning tasks, outperforming chain of thought (CoT) prompting with external feedback. We find that RCI combined with CoT performs better than either separately. Our code can be found here: https://github.com/posgnu/rci-agent.

The theory of statistical learning has focused on variational objectives expressed on functions. In this note, we discuss motivations to write similar objectives on measures, in particular to discuss out-of-distribution generalization and weakly-supervised learning. It raises a natural question: can one cast usual statistical learning results to objectives expressed on measures? Does the resulting construction lead to new algorithms of practical interest?

Shall we play a game? A single artificial intelligence can beat human players in chess, Go, poker and other games that require a variety of strategies to win. The AI, called Student of Games, was created by Google DeepMind, which says it is a step towards an artificial general intelligence capable of carrying out any task with superhuman performance. Martin Schmid, who worked at DeepMind on the AI but who is now at a start-up called EquiLibre Technologies, says that the Student of Games (SoG) model can trace its lineage back to two projects. One was DeepStack, the AI created by a team including Schmid at the University of Alberta in Canada and which was the first to beat human professional players at poker.

Microsoft wants to make its AI service for work, Microsoft 365 Copilot, basically ubiquitous within its ecosystem. And at its Ignite conference, Microsoft is well on its way to launch Copilot services early next year. For consumers, Copilot means adding additional AI capabilities to Microsoft 365 Office apps like Outlook and Teams. Microsoft is planning to remove the "Bing Chat" brand that has marked its early forays into AI and just replace it with "Copilot," for both consumers and enterprises. Copilot is also moving out of preview, and will become generally available starting December 1. If you've used Windows Copilot -- especially while working with the Copilot sidebar open -- you may have noticed Copilot react to what you're doing elsewhere on the screen.

Two days later on another stage, in another venue, at another developers' conference, Nadella made his second unannounced appearance of the week--this time at GitHub Universe. There Thomas Dohmke, GitHub's CEO, was showing off a new version of the company's AI programming tool, Copilot, that can generate computer code from natural language. Nadella was effusive: "I can code again!" he exclaimed. Today, Nadella will be onstage speaking to developers at Microsoft Ignite, where the company is announcing even more AI-based developer tools, including an Azure AI Studio that will let devs choose between model catalogs from not only Microsoft, but also the likes of Meta, OpenAI, and Hugging Face, as well as new tools for customizing Copilot for Microsoft 365. If it seems like Nadella is obsessed with developers, you're not wrong.

Microsoft has announced a project it has been "refining in secret for years;" Its own custom silicon in the form of two new server chips. The company unveiled the fruits of its labor at Microsoft Ignite, showing off the Azure Maia AI Accelerator and the Azure Cobalt CPU. The latter of which, at least, the company is happy to admit is ARM-based, which can still feel unthinkable to eyes so used to Microsoft and Intel's hand-in-glove dominance of the computing market. The company turned to OpenAI to receive feedback on Azure Maia and to use the company's models for testing. OpenAI CEO Sam Altman said the updated Microsoft's Azure will also provide the opportunity for training improved models and making them more affordable for customers.