As Deep Neural Networks (DNNs) are increasingly deployed in safety critical and privacy sensitive applications such as autonomous driving and biometric authentication, it is critical to understand the fault-tolerance nature of DNNs. Prior work primarily focuses on metrics such as Failures In Time (FIT) rate and the Silent Data Corruption (SDC) rate, which quantify how often a device fails. Instead, this paper focuses on quantifying the DNN accuracy given that a transient error has occurred, which tells us how well a network behaves when a transient error occurs. We call this metric Resiliency Accuracy (RA). We show that existing RA formulation is fundamentally inaccurate, because it incorrectly assumes that software variables (model weights/activations) have equal faulty probability under hardware transient faults. We present an algorithm that captures the faulty probabilities of DNN variables under transient faults and, thus, provides correct RA estimations validated by hardware. To accelerate RA estimation, we reformulate RA calculation as a Monte Carlo integration problem, and solve it using importance sampling driven by DNN specific heuristics. Using our lightweight RA estimation method, we show that transient faults lead to far greater accuracy degradation than what todays DNN resiliency tools estimate. We show how our RA estimation tool can help design more resilient DNNs by integrating it with a Network Architecture Search framework.

是否可以在深网络中重组非线性激活函数以创建硬件有效的模型？为了解决这个问题，我们提出了一个称为重组激活网络（RANS）的新范式，该范式操纵模型中的非线性数量以提高其硬件意识和效率。首先，我们提出了RAN-STHICER（RAN-E） - 一个新的硬件感知搜索空间和半自动搜索算法 - 用硬件感知的块替换效率低下的块。接下来，我们提出了一种称为RAN-IMPLICIC（RAN-I）的无训练模型缩放方法，从理论上讲，我们在非线性单元的数量方面证明了网络拓扑与其表现性之间的联系。我们证明，我们的网络在不同尺度和几种类型的硬件上实现最新的成像网结果。例如，与有效网络-lite-B0相比，RAN-E在ARM Micro-NPU上每秒（FPS）提高了1.5倍，同时提高了类似的精度。另一方面，ran-i以相似或更好的精度表现出#macs的#macs降低2倍。我们还表明，在基于ARM的数据中心CPU上，RAN-I的FPS比Convnext高40％。最后，与基于Convnext的模型相比，基于RAN-I的对象检测网络在数据中心CPU上获得了类似或更高的映射，并且在数据中心CPU上的fps高达33％。

自治系统对深度神经网络（DNN）的各种对抗攻击非常容易受到影响。由于其速度，易于部署以及在许多DNN上工作的能力，自由培训的模型 - 无症防御最近获得了普及。为此，已经出现了一种新技术，用于减轻对图像分类DNN的攻击，即使用超分辨率的预处理对抗性图像 - 将低质量输入提升为高分辨率图像。这种防御需要在受约束的自治系统上运行图像分类器和超分辨率模型。但是，超级分辨率招收了沉重的计算成本。因此，在本文中，我们调查以下问题：如果我们使用小型超分辨率模型，图像分类器的稳健性会受到痛苦吗？为了回答这一点，我们首先审查最近的工作称为超高效的超分辨率（SESR），其比现有技术更好地实现了类似或更好的图像质量，同时需要2x到330倍，乘法累积（MAC）操作较少。我们证明，尽管是比现有模型小的数量级，但SESR实现了与网络更大的稳健性相同。最后，我们在商业臂ETHOS-U55 Micro-NPU上估计基于超分辨率的防御的端到端性能。我们的研究结果表明，SESR在实现类似的稳健性时比基线实现了近3倍。

IOT应用中的总是关于Tinyml的感知任务需要非常高的能量效率。模拟计算内存（CIM）使用非易失性存储器（NVM）承诺高效率，并提供自包含的片上模型存储。然而，模拟CIM推出了新的实际考虑因素，包括电导漂移，读/写噪声，固定的模数转换器增益等。必须解决这些附加约束，以实现可以通过可接受的模拟CIM部署的模型精度损失。这项工作描述了$ \ textit {analognets} $：tinyml模型用于关键字点（kws）和视觉唤醒词（VWW）的流行始终是on。模型架构专门为模拟CIM设计，我们详细介绍了一种全面的培训方法，以在推理时间内保持面对模拟非理想的精度和低精度数据转换器。我们还描述了AON-CIM，可编程，最小面积的相变存储器（PCM）模拟CIM加速器，具有新颖的层串行方法，以消除与完全流水线设计相关的复杂互连的成本。我们在校准的模拟器以及真正的硬件中评估了对校准模拟器的矛盾，并发现精度下降限制为KWS / VWW的PCM漂移（8位）24小时后的0.8 $ \％$ / 1.2 $ \％$。在14nm AON-CIM加速器上运行的analognets使用8位激活，分别使用8位激活，并增加到57.39 / 25.69个顶部/ w，以4美元$ 4 $ 57.39 / 25.69。

我们提出了一种用于分布式培训神经网络模型的新型联合学习方法，其中服务器在每轮中随机选择的设备的子集之间编制协作。我们主要从通信角度查看联合学习问题，并允许更多设备级别计算来节省传输成本。我们指出了一个基本的困境，因为当地 - 设备水平的最低实证损失与全球经验损失的最小值不一致。与最近的事先有关的不同，尝试无所作用的最小化或利用用于并行化梯度计算的设备，我们为每轮的每个设备提出动态规范器，以便在极限中，全局和设备解决方案对齐。我们通过实证结果对真实的和合成数据以及我们的方案在凸和非凸面设置中导致有效培训的分析结果，同时对设备异质性完全不可知，以及大量设备，部分参与和不平衡的数据。

利用稀疏性是加速在移动设备上的量化卷积神经网络（CNN）推断的关键技术。现有稀疏的CNN加速器主要利用无结构性稀疏性并实现显着的加速。然而，由于无界，很大程度上不可预测的稀疏模式，利用非结构化稀疏性需要复杂的硬件设计，具有显着的能量和面积开销，这对能量和区域效率至关重要的移动/ IOT推理场景特别有害。我们建议利用结构化的稀疏性，更具体地，更密集地绑定块（DBB）稀疏性，用于重量和激活。 DBB块张于每个块的最大非零数。因此，DBB暴露静态可预测的稀疏模式，使瘦稀疏性利用硬件能够。我们提出了新的硬件基元，以分别为（静态）权重和（动态）激活的DBB稀疏性，具有非常低的开销。建立在基元的顶部，我们描述了一种基于收缩阵列的CNN加速器的S2TA，可利用联合重量和激活DBB稀疏性和传统的收缩系统阵列上不可用的数据重用的新维度。与具有零值时钟门控的完全阵列的强基线相比，16NM中的S2TA达到超过2倍的加速和能量减少，超过五个流行的CNN基准。与近期的非收缩稀疏加速器相比，Eyeriss V2（65nm）和Sparten（45nm），S2TA在65nm中使用约2.2倍和3.1倍的每次推断的能量较少。

The success of neural networks builds to a large extent on their ability to create internal knowledge representations from real-world high-dimensional data, such as images, sound, or text. Approaches to extract and present these representations, in order to explain the neural network's decisions, is an active and multifaceted research field. To gain a deeper understanding of a central aspect of this field, we have performed a targeted review focusing on research that aims to associate internal representations with human understandable concepts. In doing this, we added a perspective on the existing research by using primarily deductive nomological explanations as a proposed taxonomy. We find this taxonomy and theories of causality, useful for understanding what can be expected, and not expected, from neural network explanations. The analysis additionally uncovers an ambiguity in the reviewed literature related to the goal of model explainability; is it understanding the ML model or, is it actionable explanations useful in the deployment domain?

Many problems in machine learning involve bilevel optimization (BLO), including hyperparameter optimization, meta-learning, and dataset distillation. Bilevel problems consist of two nested sub-problems, called the outer and inner problems, respectively. In practice, often at least one of these sub-problems is overparameterized. In this case, there are many ways to choose among optima that achieve equivalent objective values. Inspired by recent studies of the implicit bias induced by optimization algorithms in single-level optimization, we investigate the implicit bias of gradient-based algorithms for bilevel optimization. We delineate two standard BLO methods -- cold-start and warm-start -- and show that the converged solution or long-run behavior depends to a large degree on these and other algorithmic choices, such as the hypergradient approximation. We also show that the inner solutions obtained by warm-start BLO can encode a surprising amount of information about the outer objective, even when the outer parameters are low-dimensional. We believe that implicit bias deserves as central a role in the study of bilevel optimization as it has attained in the study of single-level neural net optimization.

An expansion of aberrant brain cells is referred to as a brain tumor. The brain's architecture is extremely intricate, with several regions controlling various nervous system processes. Any portion of the brain or skull can develop a brain tumor, including the brain's protective coating, the base of the skull, the brainstem, the sinuses, the nasal cavity, and many other places. Over the past ten years, numerous developments in the field of computer-aided brain tumor diagnosis have been made. Recently, instance segmentation has attracted a lot of interest in numerous computer vision applications. It seeks to assign various IDs to various scene objects, even if they are members of the same class. Typically, a two-stage pipeline is used to perform instance segmentation. This study shows brain cancer segmentation using YOLOv5. Yolo takes dataset as picture format and corresponding text file. You Only Look Once (YOLO) is a viral and widely used algorithm. YOLO is famous for its object recognition properties. You Only Look Once (YOLO) is a popular algorithm that has gone viral. YOLO is well known for its ability to identify objects. YOLO V2, V3, V4, and V5 are some of the YOLO latest versions that experts have published in recent years. Early brain tumor detection is one of the most important jobs that neurologists and radiologists have. However, it can be difficult and error-prone to manually identify and segment brain tumors from Magnetic Resonance Imaging (MRI) data. For making an early diagnosis of the condition, an automated brain tumor detection system is necessary. The model of the research paper has three classes. They are respectively Meningioma, Pituitary, Glioma. The results show that, our model achieves competitive accuracy, in terms of runtime usage of M2 10 core GPU.

Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation, but the quality bar for medical and clinical applications is high. Today, attempts to assess models' clinical knowledge typically rely on automated evaluations on limited benchmarks. There is no standard to evaluate model predictions and reasoning across a breadth of tasks. To address this, we present MultiMedQA, a benchmark combining six existing open question answering datasets spanning professional medical exams, research, and consumer queries; and HealthSearchQA, a new free-response dataset of medical questions searched online. We propose a framework for human evaluation of model answers along multiple axes including factuality, precision, possible harm, and bias. In addition, we evaluate PaLM (a 540-billion parameter LLM) and its instruction-tuned variant, Flan-PaLM, on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA, MedMCQA, PubMedQA, MMLU clinical topics), including 67.6% accuracy on MedQA (US Medical License Exam questions), surpassing prior state-of-the-art by over 17%. However, human evaluation reveals key gaps in Flan-PaLM responses. To resolve this we introduce instruction prompt tuning, a parameter-efficient approach for aligning LLMs to new domains using a few exemplars. The resulting model, Med-PaLM, performs encouragingly, but remains inferior to clinicians. We show that comprehension, recall of knowledge, and medical reasoning improve with model scale and instruction prompt tuning, suggesting the potential utility of LLMs in medicine. Our human evaluations reveal important limitations of today's models, reinforcing the importance of both evaluation frameworks and method development in creating safe, helpful LLM models for clinical applications.