预测风险评分越来越多地用于指导复杂环境（尤其是医疗保健）中的临床或其他干预措施。直接更新用于指导干预措施的风险评分会导致风险估计。我们建议使用“保留集”（未接受风险评分引导干预措施的人口子集）进行更新，以防止这种情况。由于保留集中的样本并不能从风险预测中受益，因此其规模必须权衡更新的风险评分的性能，同时最大程度地减少被保留样品的数量。我们证明，这种方法的表现优于简单的替代方案，并且通过定义一般的损失函数描述了可以轻松识别最佳保持尺寸（OHS）的条件。我们引入了OHS估计的参数和半参数算法，并证明了它们在近期对先兆子痫的风险评分上的使用。基于这些结果，我们认为保留集是安全，可行且易于实施的手段，可以安全地更新预测风险得分。

Naturally-occurring information-seeking questions often contain questionable assumptions -- assumptions that are false or unverifiable. Questions containing questionable assumptions are challenging because they require a distinct answer strategy that deviates from typical answers to information-seeking questions. For instance, the question "When did Marie Curie discover Uranium?" cannot be answered as a typical when question without addressing the false assumption "Marie Curie discovered Uranium". In this work, we propose (QA)$^2$ (Question Answering with Questionable Assumptions), an open-domain evaluation dataset consisting of naturally-occurring search engine queries that may or may not contain questionable assumptions. To be successful on (QA)$^2$, systems must be able to detect questionable assumptions and also be able to produce adequate responses for both typical information-seeking questions and ones with questionable assumptions. We find that current models do struggle with handling questionable assumptions -- the best performing model achieves 59% human rater acceptability on abstractive QA with (QA)$^2$ questions, leaving substantial headroom for progress.

As language models (LMs) scale, they develop many novel behaviors, good and bad, exacerbating the need to evaluate how they behave. Prior work creates evaluations with crowdwork (which is time-consuming and expensive) or existing data sources (which are not always available). Here, we automatically generate evaluations with LMs. We explore approaches with varying amounts of human effort, from instructing LMs to write yes/no questions to making complex Winogender schemas with multiple stages of LM-based generation and filtering. Crowdworkers rate the examples as highly relevant and agree with 90-100% of labels, sometimes more so than corresponding human-written datasets. We generate 154 datasets and discover new cases of inverse scaling where LMs get worse with size. Larger LMs repeat back a dialog user's preferred answer ("sycophancy") and express greater desire to pursue concerning goals like resource acquisition and goal preservation. We also find some of the first examples of inverse scaling in RL from Human Feedback (RLHF), where more RLHF makes LMs worse. For example, RLHF makes LMs express stronger political views (on gun rights and immigration) and a greater desire to avoid shut down. Overall, LM-written evaluations are high-quality and let us quickly discover many novel LM behaviors.

As AI systems become more capable, we would like to enlist their help to supervise other AIs. We experiment with methods for training a harmless AI assistant through self-improvement, without any human labels identifying harmful outputs. The only human oversight is provided through a list of rules or principles, and so we refer to the method as 'Constitutional AI'. The process involves both a supervised learning and a reinforcement learning phase. In the supervised phase we sample from an initial model, then generate self-critiques and revisions, and then finetune the original model on revised responses. In the RL phase, we sample from the finetuned model, use a model to evaluate which of the two samples is better, and then train a preference model from this dataset of AI preferences. We then train with RL using the preference model as the reward signal, i.e. we use 'RL from AI Feedback' (RLAIF). As a result we are able to train a harmless but non-evasive AI assistant that engages with harmful queries by explaining its objections to them. Both the SL and RL methods can leverage chain-of-thought style reasoning to improve the human-judged performance and transparency of AI decision making. These methods make it possible to control AI behavior more precisely and with far fewer human labels.

Machine Learning models capable of handling the large datasets collected in the financial world can often become black boxes expensive to run. The quantum computing paradigm suggests new optimization techniques, that combined with classical algorithms, may deliver competitive, faster and more interpretable models. In this work we propose a quantum-enhanced machine learning solution for the prediction of credit rating downgrades, also known as fallen-angels forecasting in the financial risk management field. We implement this solution on a neutral atom Quantum Processing Unit with up to 60 qubits on a real-life dataset. We report competitive performances against the state-of-the-art Random Forest benchmark whilst our model achieves better interpretability and comparable training times. We examine how to improve performance in the near-term validating our ideas with Tensor Networks-based numerical simulations.

Developing safe and useful general-purpose AI systems will require us to make progress on scalable oversight: the problem of supervising systems that potentially outperform us on most skills relevant to the task at hand. Empirical work on this problem is not straightforward, since we do not yet have systems that broadly exceed our abilities. This paper discusses one of the major ways we think about this problem, with a focus on how to turn it into one that can be productively studied empirically. We first present an experimental design centered on choosing tasks for which human specialists succeed but unaided humans and current general AI systems fail. We then present a proof-of-concept experiment following meant to demonstrate a key feature of this experimental design and show its viability with two question-answering tasks: MMLU and time-limited QuALITY. On these tasks, we find that human participants who interact with an unreliable large-language-model dialog assistant through chat -- a trivial baseline strategy for scalable oversight -- substantially outperform both the model alone and their own unaided performance. These results are an encouraging sign that scalable oversight will be tractable to study with present models and bolster recent findings that large language models can productively assist humans with difficult tasks.

在这项研究中，将放射学方法扩展到用于组织分类的光学荧光分子成像数据，称为“验光”。荧光分子成像正在出现在头颈部鳞状细胞癌（HNSCC）切除期间的精确手术引导。然而，肿瘤到正常的组织对比与靶分子表皮生长因子受体（EGFR）的异质表达的内在生理局限性混淆。验光学试图通过探测荧光传达的EGFR表达中的质地模式差异来改善肿瘤识别。从荧光图像样品中提取了总共1,472个标准化的验光特征。涉及支持矢量机分类器的监督机器学习管道接受了25个顶级功能的培训，这些功能由最小冗余最大相关标准选择。通过将切除组织的图像贴片分类为组织学确认的恶性肿瘤状态，将模型预测性能与荧光强度阈值方法进行了比较。与荧光强度阈值方法相比，验光方法在所有测试集样品中提供了一致的预测准确性（无剂量）（平均精度为89％vs. 81％； P = 0.0072）。改进的性能表明，将放射线学方法扩展到荧光分子成像数据为荧光引导手术中的癌症检测提供了有希望的图像分析技术。

我们介绍了NLP社区Metasurvey的结果。从2022年5月到2022年6月，该调查引起了关于有争议的问题的意见，包括该领域的行业影响，对AGI和道德规范的关注。我们的结果将具体数字置于几个争议中：例如，受访者几乎完全将有关人工通用智能的重要性的问题分为一半，语言模型是否理解语言以及语言结构的必要性以及解决NLP问题的必要性。此外，调查提出了元问题，要求受访者预测调查响应的分布。这不仅使我们不仅可以深入了解NLP研究人员所拥有的各种信念，还可以揭示社区预测与现实不符的错误社会学信念。我们在各种问题上发现这种不匹配。除其他结果外，社区大大高估了其对基准的实用性的信念，以及扩展解决现实世界中问题的潜力，同时低估了其对语言结构，归纳偏见和跨学科科学重要性的信念。

自然语言处理的机器学习快速进步有可能改变有关人类学习语言的辩论。但是，当前人工学习者和人类的学习环境和偏见以削弱从学习模拟获得的证据的影响的方式分歧。例如，当今最有效的神经语言模型接受了典型儿童可用的语言数据量的大约一千倍。为了增加计算模型的可学习性结果的相关性，我们需要培训模型学习者，而没有比人类具有显着优势的学习者。如果合适的模型成功地获得了一些目标语言知识，则可以提供一个概念证明，即在假设的人类学习方案中可以学习目标。合理的模型学习者将使我们能够进行实验操作，以对学习环境中的变量进行因果推断，并严格测试史密斯风格的贫困声明，主张根据人类对人类的先天语言知识，基于有关可学习性的猜测。由于实用和道德的考虑因素，人类受试者将永远无法实现可比的实验，从而使模型学习者成为必不可少的资源。到目前为止，试图剥夺当前模型的不公平优势，为关键语法行为（例如可接受性判断）获得亚人类结果。但是，在我们可以合理地得出结论，语言学习需要比当前模型拥有更多的特定领域知识，我们必须首先以多模式刺激和多代理互动的形式探索非语言意见，以使学习者更有效地学习学习者来自有限的语言输入。

脑病理通常表现为组织的部分或完全丧失。许多神经影像学研究的目的是捕获有关感兴趣的临床变量（例如疾病进展）的组织变化的位置和数量。形态分析方法捕获了与临床变量有关的组织分布或其他含量的兴趣分布的局部差异。我们建议通过基于不平衡的最佳传输的附加特征提取步骤来增强形态分析。最佳运输特征提取步骤增加了导致空间分散组织损失的病理学的统计能力，从而最大程度地减少了由于空间未对准或大脑拓扑差异而对变化的敏感性，并将由于组织位置而导致的变化而分离。我们证明了在阿尔茨海默氏病的OASIS-1研究的体积形态学分析的背景下，提出的最佳运输特征提取步骤。结果表明，所提出的方法可以识别组织的变化和差异，而这些差异是无法测量的。