Code completion is essential in software development, helping developers by predicting code snippets based on context. Among completion tasks, Method Body Completion (MBC) is particularly challenging as it involves generating complete method bodies based on their signatures and context. This task becomes significantly harder in large repositories, where method bodies must integrate repositoryspecific elements such as custom APIs, inter-module dependencies, and project-specific conventions. In this paper, we introduce RAMBO, a novel RAG-based approach for repository-level MBC. Instead of retrieving similar method bodies, RAMBO identifies essential repository-specific elements, such as classes, methods, and variables/fields, and their relevant usages. By incorporating these elements and their relevant usages into the code generation process, RAMBO ensures more accurate and contextually relevant method bodies. Our experimental results with leading code LLMs across 40 Java projects show that RAMBO significantly outperformed the state-of-the-art repository-level MBC approaches, with the improvements of up to 46% in BLEU, 57% in CodeBLEU, 36% in Compilation Rate, and up to 3X in Exact Match. Notably, RAMBO surpassed RepoCoder Oracle method by up to 12% in Exact Match, setting a new benchmark for repository-level MBC.

Important variables of processes are, in many occasions, categorical, i.e. names or labels representing, e.g. categories of inputs, or types of reactors or a sequence of steps. In this work, we use Large Language Models (LLMs) to derive embeddings of such inputs that represent their actual meaning, or reflect the ``distances" between categories, i.e. how similar or dissimilar they are. This is a marked difference from the current standard practice of using binary, or one-hot encoding to replace categorical variables with sequences of ones and zeros. Combined with dimensionality reduction techniques, either linear such as Principal Components Analysis (PCA), or nonlinear such as Uniform Manifold Approximation and Projection (UMAP), the proposed approach leads to a \textit{meaningful}, low-dimensional feature space. The significance of obtaining meaningful embeddings is illustrated in the context of an industrial coating process for cutting tools that includes both numerical and categorical inputs. The proposed approach enables feature importance which is a marked improvement compared to the current state-of-the-art (SotA) in the encoding of categorical variables.

The increasing administrative burden of medical documentation, particularly through Electronic Health Records (EHR), significantly reduces the time available for direct patient care and contributes to physician burnout. To address this issue, we propose MediNotes, an advanced generative AI framework designed to automate the creation of SOAP (Subjective, Objective, Assessment, Plan) notes from medical conversations. MediNotes integrates Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), and Automatic Speech Recognition (ASR) to capture and process both text and voice inputs in real time or from recorded audio, generating structured and contextually accurate medical notes. The framework also incorporates advanced techniques like Quantized Low-Rank Adaptation (QLoRA) and Parameter-Efficient Fine-Tuning (PEFT) for efficient model fine-tuning in resource-constrained environments. Additionally, MediNotes offers a query-based retrieval system, allowing healthcare providers and patients to access relevant medical information quickly and accurately. Evaluations using the ACI-BENCH dataset demonstrate that MediNotes significantly improves the accuracy, efficiency, and usability of automated medical documentation, offering a robust solution to reduce the administrative burden on healthcare professionals while improving the quality of clinical workflows.

OpenAI, the developer of the groundbreaking ChatGPT chatbot, is preparing to overhaul its corporate structure and become a for-profit business. The startup's chief executive, Sam Altman, acknowledged on Thursday that it was "not a normal company" after another surprising development at OpenAI this week when its its chief technology officer, Mira Murati, resigned. Her departure was quickly followed by the announcement that two other executives had quit. The company is synonymous with an artificial intelligence boom triggered by the emergence, in 2022, of OpenAI's signature product, a chatbot that stunned users with its ability to craft convincing, human-like responses to an array of prompts. Altman, in turn, has become the poster child for a technology that is advancing rapidly and is being developed by the world's largest tech companies, including Microsoft – OpenAI's biggest backer – Google, the Facebook owner Meta and Amazon.

The update means Google's Gemini 1.5 Pro-powered virtual assistant can now handle Google Docs, PDFs, text files, Google Slides, YouTube video URLs, audio files and web pages. This expands NotebookLM's potential sources to include lecture recordings, informative YouTube content and group discussions. Google claims that it securely stores all information uploaded to NoteBookLM and does not use it to train AI. Google launched the virtual research assistant in the summer of 2023 and upgraded it to run on Gemini 1.5 Pro this past June. Alongside the AI-powered bump, Google expanded NotebookLM to be available in over 200 countries and territories.

There's a story about Sam Altman that has been repeated often enough to become Silicon Valley lore. In 2012, Paul Graham, a co-founder of the famed start-up accelerator Y Combinator and one of Altman's biggest mentors, sat Altman down and asked if he wanted to take over the organization. The decision was a peculiar one: Altman was only in his late 20s, and at least on paper, his qualifications were middling. He had dropped out of Stanford to found a company that ultimately hadn't panned out. After seven years, he'd sold it for roughly the same amount that his investors had put in.

In the decade-long fight to control CRISPR, the super-tool for modifying DNA, it's been common for lawyers to try to overturn patents held by competitors. But now, in a surprise twist, the team that earned the Nobel Prize in chemistry for developing CRISPR is asking to cancel two of their own seminal patents, MIT Technology Review has learned. The request to withdraw the pair of European patents, by lawyers for Emmanuelle Charpentier and Jennifer Doudna, comes after a damaging August opinion from a European technical appeals board, which ruled that the duo's earliest patent filing didn't explain CRISPR well enough for other scientists to use it and doesn't count as a proper invention. The decision could have major ramifications regarding who gets to collect the lucrative licensing fees on using the technology.Read the full story. What's new: The Allen Institute for Artificial Intelligence (Ai2), a research nonprofit, is releasing a family of open-source multimodal language models, called Molmo, that it says perform as well as top proprietary models from OpenAI, Google, and Anthropic.

OpenAI is reportedly pushing ahead with plans to become a for-profit company, as more senior figures left the ChatGPT developer after the surprise exit of its chief technology officer, Mira Murati. The San Francisco-based startup is preparing to change its corporate structure as it seeks 6.5bn ( 4.9bn) of new funding, according to reports. Under the changes, OpenAI will become a for-profit benefit corporation – an entity that makes profits but is committed to the social and public good – that will no longer be controlled by its nonprofit board, Reuters reported. OpenAI declined to comment on the details of the reports but a spokesperson said the nonprofit entity would continue to exist. "We remain focused on building AI that benefits everyone, and we're working with our board to ensure that we're best positioned to succeed in our mission. The nonprofit is core to our mission and will continue to exist," said the spokesperson.

A high-ranking executive at OpenAI who served a few days as its interim CEO during a period of turmoil last year said she's leaving the artificial intelligence company. Mira Murati, OpenAI's chief technology officer, said in a written statement Wednesday that, after much reflection, she has "made the difficult decision to leave OpenAI." "I'm stepping away because I want to create the time and space to do my own exploration," she said. Two other top executives are also on their way out, CEO Sam Altman announced later Wednesday. The decisions by Murati, as well as OpenAI's Chief Research Officer Bob McGrew and another research leader, Barret Zoph, were made "independently of each other and amicably," Altman said in a note to employees he shared on social media.

Recently, there has been significant interest in replacing the reward model in Reinforcement Learning with Human Feedback (RLHF) methods for Large Language Models (LLMs), such as Direct Preference Optimization (DPO) and its variants. These approaches commonly use a binary cross-entropy mechanism on pairwise samples, i.e., minimizing and maximizing the loss based on preferred or dis-preferred responses, respectively. However, while this training strategy omits the reward model, it also overlooks the varying preference degrees within different responses. We hypothesize that this is a key factor hindering LLMs from sufficiently understanding human preferences. To address this problem, we propose a novel Self-supervised Preference Optimization (SPO) framework, which constructs a self-supervised preference degree loss combined with the alignment loss, thereby helping LLMs improve their ability to understand the degree of preference. Extensive experiments are conducted on two widely used datasets of different tasks. The results demonstrate that SPO can be seamlessly integrated with existing preference optimization methods and significantly boost their performance to achieve state-of-the-art performance. We also conduct detailed analyses to offer comprehensive insights into SPO, which verifies its effectiveness. The code is available at https://github.com/lijian16/SPO.