Modern AI agents, driven by advances in large foundation models, promise to enhance our productivity and transform our lives by augmenting our knowledge and capabilities. To achieve this vision, AI agents must effectively plan, perform multi-step reasoning and actions, respond to novel observations, and recover from errors, to successfully complete complex tasks across a wide range of scenarios. In this work, we introduce Magentic-One, a high-performing open-source agentic system for solving such tasks. Magentic-One uses a multi-agent architecture where a lead agent, the Orchestrator, plans, tracks progress, and re-plans to recover from errors. Throughout task execution, the Orchestrator directs other specialized agents to perform tasks as needed, such as operating a web browser, navigating local files, or writing and executing Python code. We show that Magentic-One achieves statistically competitive performance to the state-of-the-art on three diverse and challenging agentic benchmarks: GAIA, AssistantBench, and WebArena. Magentic-One achieves these results without modification to core agent capabilities or to how they collaborate, demonstrating progress towards generalist agentic systems. Moreover, Magentic-One's modular design allows agents to be added or removed from the team without additional prompt tuning or training, easing development and making it extensible to future scenarios. We provide an open-source implementation of Magentic-One, and we include AutoGenBench, a standalone tool for agentic evaluation. AutoGenBench provides built-in controls for repetition and isolation to run agentic benchmarks in a rigorous and contained manner -- which is important when agents' actions have side-effects. Magentic-One, AutoGenBench and detailed empirical performance evaluations of Magentic-One, including ablations and error analysis are available at https://aka.ms/magentic-one

We propose a novel approach to sequential Bayesian inference based on variational Bayes. The key insight is that, in the online setting, we do not need to add the KL term to regularize to the prior (which comes from the posterior at the previous timestep); instead we can optimize just the expected log-likelihood, performing a single step of natural gradient descent starting at the prior predictive. We prove this method recovers exact Bayesian inference if the model is conjugate, and empirically outperforms other online VB methods in the non-conjugate setting, such as online learning for neural networks, especially when controlling for computational costs.

Estimation of heterogeneous treatment effects from observational data has become an important problem. It plays a crucial role in determining the individualized causal effects of a treatment, which then leads to a personalized assignment of optimal treatment (Wendling et al., 2018; Rekkas et al., 2020). Estimation of such heterogeneity however requires reasonable representations from each treatment subgroup. With the increasing availability of large-scale health outcome data such as electronic health records (EHR) data in recent years, it has become possible to develop individualized treatment strategies efficiently. This led to the development of several novel statistical methods, primarily tailored for binary treatment scenarios (Wendling et al., 2018; Cheng et al., 2020), with some accommodating multiple treatment settings (Brown et al., 2020; Chalkou et al., 2021). Most of these approaches are specifically designed for estimating population average treatment effects (ATEs) (Van Der Laan and Rubin, 2006; Chernozhukov et al., 2018; McCaffrey et al., 2013) and more recently, methods are being developed to estimate conditional average treatment effects (CATEs) (Taddy et al., 2016; Wager and Athey, 2018; Künzel et al., 2019; Nie and Wager, 2021). Here, we tackle a generic problem of heterogeneous treatment effect or CATE estimation in a multi-treatment setting, where the treatment responses may share some commonalities.

This is the first survey of the active area of AI research that focuses on privacy issues in Large Language Models (LLMs). Specifically, we focus on work that red-teams models to highlight privacy risks, attempts to build privacy into the training or inference process, enables efficient data deletion from trained models to comply with existing privacy regulations, and tries to mitigate copyright issues. Our focus is on summarizing technical research that develops algorithms, proves theorems, and runs empirical evaluations. While there is an extensive body of legal and policy work addressing these challenges from a different angle, that is not the focus of our survey. Nevertheless, these works, along with recent legal developments do inform how these technical problems are formalized, and so we discuss them briefly in Section 1. While we have made our best effort to include all the relevant work, due to the fast moving nature of this research we may have missed some recent work. If we have missed some of your work please contact us, as we will attempt to keep this survey relatively up to date. We are maintaining a repository with the list of papers covered in this survey and any relevant code that was publicly available at https://github.com/safr-ml-lab/survey-llm.

Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabilities are as yet poorly characterized. In order to inform future research, prepare for disruptive new model capabilities, and ameliorate socially harmful effects, it is vital that we understand the present and near-future capabilities and limitations of language models. To address this challenge, we introduce the Beyond the Imitation Game benchmark (BIG-bench). BIG-bench currently consists of 204 tasks, contributed by 450 authors across 132 institutions. Task topics are diverse, drawing problems from linguistics, childhood development, math, common-sense reasoning, biology, physics, social bias, software development, and beyond. BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models. We evaluate the behavior of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters. In addition, a team of human expert raters performed all tasks in order to provide a strong baseline. Findings include: model performance and calibration both improve with scale, but are poor in absolute terms (and when compared with rater performance); performance is remarkably similar across model classes, though with benefits from sparsity; tasks that improve gradually and predictably commonly involve a large knowledge or memorization component, whereas tasks that exhibit "breakthrough" behavior at a critical scale often involve multiple steps or components, or brittle metrics; social bias typically increases with scale in settings with ambiguous context, but this can be improved with prompting.