OpenAI just launched its AI search update for ChatGPT. Three months after the company's initial announcement of a SearchGPT prototype, OpenAI's vision for the future of AI search is now available to the public. "We're focused on making ChatGPT the best place to answer any question, including live information from the web," says Adam Fry, the product lead for search on ChatGPT. Referred to by Fry now as "ChatGPT search" rather than "SearchGPT," the feature enters an increasingly crowded and contentious field of AI search options for users--with competition from smaller startups, like Perplexity, as well as tech giants, like Google with its AI Overview search results. So far in 2024, journalists have criticized both Google and Perplexity's implementations of AI search for improperly copying aspects of original work and hallucinating fake information.

"Our goal is to make ChatGPT the smartest assistant, and now we're really enhancing its capabilities in terms of what it has access to from the web," Fry tells MIT Technology Review. The feature is available today for the chatbot's paying users. While ChatGPT search, as it is known, is initially available to paying customers, OpenAI intends to make it available for free later, even when people are logged out. The company also plans to combine search with its voice features and Canvas, its interactive platform for coding and writing, although these capabilities will not be available in today's initial launch. The company unveiled a standalone prototype of web search in July.

Meta CEO Mark Zuckerberg laid down the newest marker in generative AI training on Wednesday, saying that the next major release of the company's Llama model is being trained on a cluster of GPUs that's "bigger than anything" else that's been reported. Llama 4 development is well underway, Zuckerberg told investors and analysts on an earnings call, with an initial launch expected early next year. "We're training the Llama 4 models on a cluster that is bigger than 100,000 H100s, or bigger than anything that I've seen reported for what others are doing," Zuckerberg said, referring to the Nvidia chips popular for training AI systems. "I expect that the smaller Llama 4 models will be ready first." Increasing the scale of AI training with more computing power and data is widely believed to be key to developing significantly more capable AI models.

Large Language Models (LLMs) have transformed how people interact with artificial intelligence (AI) systems, achieving state-of-the-art results in various tasks, including scientific discovery and hypothesis generation. However, the lack of a comprehensive and systematic evaluation framework for generating research ideas using LLMs poses a significant obstacle to understanding and assessing their generative capabilities in scientific discovery. To address this gap, we propose IdeaBench, a benchmark system that includes a comprehensive dataset and an evaluation framework for standardizing the assessment of research idea generation using LLMs. Our dataset comprises titles and abstracts from a diverse range of influential papers, along with their referenced works. To emulate the human process of generating research ideas, we profile LLMs as domain-specific researchers and ground them in the same context considered by human researchers. This maximizes the utilization of the LLMs' parametric knowledge to dynamically generate new research ideas. We also introduce an evaluation framework for assessing the quality of generated research ideas. Our evaluation framework is a two-stage process: first, using GPT-4o to rank ideas based on user-specified quality indicators such as novelty and feasibility, enabling scalable personalization; and second, calculating relative ranking based "Insight Score" to quantify the chosen quality indicator. The proposed benchmark system will be a valuable asset for the community to measure and compare different LLMs, ultimately advancing the automation of the scientific discovery process.

Recent advances in aligning Large Language Models with human preferences have benefited from larger reward models and better preference data. However, most of these methodologies rely on the accuracy of the reward model. The reward models used in Reinforcement Learning with Human Feedback (RLHF) are typically learned from small datasets using stochastic optimization algorithms, making them prone to high variability. We illustrate the inconsistencies between reward models empirically on numerous open-source datasets. We theoretically show that the fluctuation of the reward models can be detrimental to the alignment problem because the derived policies are more overfitted to the reward model and, hence, are riskier if the reward model itself is uncertain. We use concentration of measure to motivate an uncertainty-aware, conservative algorithm for policy optimization. We show that such policies are more risk-averse in the sense that they are more cautious of uncertain rewards. We theoretically prove that our proposed methodology has less risk than the vanilla method. We corroborate our theoretical results with experiments based on designing an ensemble of reward models. We use this ensemble of reward models to align a language model using our methodology and observe that our empirical findings match our theoretical predictions.

Due to GPT's impressive generative capabilities, its applications in games are expanding rapidly. To offer researchers a comprehensive understanding of the current applications and identify both emerging trends and unexplored areas, this paper introduces an updated scoping review of 131 articles, 76 of which were published in 2024, to explore GPT's potential for games. By coding and synthesizing the papers, we identify five prominent applications of GPT in current game research: procedural content generation, mixed-initiative game design, mixed-initiative gameplay, playing games, and game user research. Drawing on insights from these application areas and emerging research, we propose future studies should focus on expanding the technical boundaries of the GPT models and exploring the complex interaction dynamics between them and users. This review aims to illustrate the state of the art in innovative GPT applications in games, offering a foundation to enrich game development and enhance player experiences through cutting-edge AI innovations.

Differential diagnosis is crucial for medicine as it helps healthcare providers systematically distinguish between conditions that share similar symptoms. This study assesses the impact of lab test results on differential diagnoses (DDx) made by large language models (LLMs). Clinical vignettes from 50 case reports from PubMed Central were created incorporating patient demographics, symptoms, and lab results. Five LLMs GPT-4, GPT-3.5, Llama-2-70b, Claude-2, and Mixtral-8x7B were tested to generate Top 10, Top 5, and Top 1 DDx with and without lab data. A comprehensive evaluation involving GPT-4, a knowledge graph, and clinicians was conducted. GPT-4 performed best, achieving 55% accuracy for Top 1 diagnoses and 60% for Top 10 with lab data, with lenient accuracy up to 80%. Lab results significantly improved accuracy, with GPT-4 and Mixtral excelling, though exact match rates were low. Lab tests, including liver function, metabolic/toxicology panels, and serology/immune tests, were generally interpreted correctly by LLMs for differential diagnosis.

This paper presents "Biotic Browser," an innovative AI assistant leveraging StreamingLLM to transform web navigation and task execution. Characterized by its ability to simulate the experience of a passenger in an autonomous vehicle, the Biotic Browser excels in managing extended interactions and complex, multi-step web-based tasks. It marks a significant advancement in AI technology, particularly in the realm of long-term context management, and offers promising applications for enhancing productivity and efficiency in both personal and professional settings.

We study the optimal memorization capacity of modern Hopfield models and Kernelized Hopfield Models (KHMs), a transformer-compatible class of Dense Associative Memories. We present a tight analysis by establishing a connection between the memory configuration of KHMs and spherical codes from information theory. Specifically, we treat the stored memory set as a specialized spherical code. This enables us to cast the memorization problem in KHMs into a point arrangement problem on a hypersphere. We show that the optimal capacity of KHMs occurs when the feature space allows memories to form an optimal spherical code. This unique perspective leads to: (i) An analysis of how KHMs achieve optimal memory capacity, and identify corresponding necessary conditions. Importantly, we establish an upper capacity bound that matches the well-known exponential lower bound in the literature. This provides the first tight and optimal asymptotic memory capacity for modern Hopfield models. (ii) A sub-linear time algorithm $\mathtt{U}\text{-}\mathtt{Hop}$+ to reach KHMs' optimal capacity. (iii) An analysis of the scaling behavior of the required feature dimension relative to the number of stored memories. These efforts improve both the retrieval capability of KHMs and the representation learning of corresponding transformers. Experimentally, we provide thorough numerical results to back up theoretical findings.

Zero-shot domain adaptation for dialogue state tracking (DST) remains a challenging problem in task-oriented dialogue (TOD) systems, where models must generalize to target domains unseen at training time. Current large language model approaches for zero-shot domain adaptation rely on prompting to introduce knowledge pertaining to the target domains. However, their efficacy strongly depends on prompt engineering, as well as the zero-shot ability of the underlying language model. In this work, we devise a novel data augmentation approach, Schema Augmentation, that improves the zero-shot domain adaptation of language models through fine-tuning. Schema Augmentation is a simple but effective technique that enhances generalization by introducing variations of slot names within the schema provided in the prompt. Experiments on MultiWOZ and SpokenWOZ showed that the proposed approach resulted in a substantial improvement over the baseline, in some experiments achieving over a twofold accuracy gain over unseen domains while maintaining equal or superior performance over all domains.