FP8是加速深度学习训练推论以外的16位格式的自然发展。在本文中，我们提出了一个8位浮点（FP8）二进制互换格式，该格式由两个编码组成-E4M3（4位指数和3位Mantissa）和E5M2（5位指数和2位指数和2位Mantissa）。尽管E5M2遵循IEEE 754惯例代表特殊值的惯例，但E4M3的动态范围是通过不代表无限态，只有一个Mantissa Bit-Pattern来扩展NAN。我们证明了FP8格式对各种图像和语言任务的功效，从而有效地匹配了16位培训课程所达到的质量。我们的研究涵盖了主要的现代神经网络体系结构 - CNN，RNN和基于变压器的模型，使所有超参数与16位基线训练课程保持不变。我们的培训实验包括大型，最多175b参数，语言模型。我们还检查了使用16位格式训练的语言模型的FP8训练后定量化，该格式抗拒固定点INT8量化。

In multi-agent systems with large number of agents, typically the contribution of each agent to the value of other agents is minimal (e.g., aggregation systems such as Uber, Deliveroo). In this paper, we consider such multi-agent systems where each agent is self-interested and takes a sequence of decisions and represent them as a Stochastic Non-atomic Congestion Game (SNCG). We derive key properties for equilibrium solutions in SNCG model with non-atomic and also nearly non-atomic agents. With those key equilibrium properties, we provide a novel Multi-Agent Reinforcement Learning (MARL) mechanism that minimizes variance across values of agents in the same state. To demonstrate the utility of this new mechanism, we provide detailed results on a real-world taxi dataset and also a generic simulator for aggregation systems. We show that our approach reduces the variance in revenues earned by taxi drivers, while still providing higher joint revenues than leading approaches.

Reduced system dependability and higher maintenance costs may be the consequence of poor electric power quality, which can disturb normal equipment performance, speed up aging, and even cause outright failures. This study implements and tests a prototype of an Online Sequential Extreme Learning Machine (OS-ELM) classifier based on wavelets for detecting power quality problems under transient conditions. In order to create the classifier, the OSELM-network model and the discrete wavelet transform (DWT) method are combined. First, discrete wavelet transform (DWT) multi-resolution analysis (MRA) was used to extract characteristics of the distorted signal at various resolutions. The OSELM then sorts the retrieved data by transient duration and energy features to determine the kind of disturbance. The suggested approach requires less memory space and processing time since it can minimize a large quantity of the distorted signal's characteristics without changing the signal's original quality. Several types of transient events were used to demonstrate the classifier's ability to detect and categorize various types of power disturbances, including sags, swells, momentary interruptions, oscillatory transients, harmonics, notches, spikes, flickers, sag swell, sag mi, sag harm, swell trans, sag spike, and swell spike.

Concept bottleneck models (CBMs) (Koh et al. 2020) are interpretable neural networks that first predict labels for human-interpretable concepts relevant to the prediction task, and then predict the final label based on the concept label predictions.We extend CBMs to interactive prediction settings where the model can query a human collaborator for the label to some concepts. We develop an interaction policy that, at prediction time, chooses which concepts to request a label for so as to maximally improve the final prediction. We demonstrate thata simple policy combining concept prediction uncertainty and influence of the concept on the final prediction achieves strong performance and outperforms a static approach proposed in Koh et al. (2020) as well as active feature acquisition methods proposed in the literature. We show that the interactiveCBM can achieve accuracy gains of 5-10% with only 5 interactions over competitive baselines on the Caltech-UCSDBirds, CheXpert and OAI datasets.

Selective classification involves identifying the subset of test samples that a model can classify with high accuracy, and is important for applications such as automated medical diagnosis. We argue that this capability of identifying uncertain samples is valuable for training classifiers as well, with the aim of building more accurate classifiers. We unify these dual roles by training a single auxiliary meta-network to output an importance weight as a function of the instance. This measure is used at train time to reweight training data, and at test-time to rank test instances for selective classification. A second, key component of our proposal is the meta-objective of minimizing dropout variance (the variance of classifier output when subjected to random weight dropout) for training the metanetwork. We train the classifier together with its metanetwork using a nested objective of minimizing classifier loss on training data and meta-loss on a separate meta-training dataset. We outperform current state-of-the-art on selective classification by substantial margins--for instance, upto 1.9% AUC and 2% accuracy on a real-world diabetic retinopathy dataset. Finally, our meta-learning framework extends naturally to unsupervised domain adaptation, given our unsupervised variance minimization meta-objective. We show cumulative absolute gains of 3.4% / 3.3% accuracy and AUC over the other baselines in domain shift settings on the Retinopathy dataset using unsupervised domain adaptation.

Many real-world learning scenarios face the challenge of slow concept drift, where data distributions change gradually over time. In this setting, we pose the problem of learning temporally sensitive importance weights for training data, in order to optimize predictive accuracy. We propose a class of temporal reweighting functions that can capture multiple timescales of change in the data, as well as instance-specific characteristics. We formulate a bi-level optimization criterion, and an associated meta-learning algorithm, by which these weights can be learned. In particular, our formulation trains an auxiliary network to output weights as a function of training instances, thereby compactly representing the instance weights. We validate our temporal reweighting scheme on a large real-world dataset of 39M images spread over a 9 year period. Our extensive experiments demonstrate the necessity of instance-based temporal reweighting in the dataset, and achieve significant improvements to classical batch-learning approaches. Further, our proposal easily generalizes to a streaming setting and shows significant gains compared to recent continual learning methods.

Models trained via empirical risk minimization (ERM) are known to rely on spurious correlations between labels and task-independent input features, resulting in poor generalization to distributional shifts. Group distributionally robust optimization (G-DRO) can alleviate this problem by minimizing the worst-case loss over a set of pre-defined groups over training data. G-DRO successfully improves performance of the worst-group, where the correlation does not hold. However, G-DRO assumes that the spurious correlations and associated worst groups are known in advance, making it challenging to apply it to new tasks with potentially multiple unknown spurious correlations. We propose AGRO -- Adversarial Group discovery for Distributionally Robust Optimization -- an end-to-end approach that jointly identifies error-prone groups and improves accuracy on them. AGRO equips G-DRO with an adversarial slicing model to find a group assignment for training examples which maximizes worst-case loss over the discovered groups. On the WILDS benchmark, AGRO results in 8% higher model performance on average on known worst-groups, compared to prior group discovery approaches used with G-DRO. AGRO also improves out-of-distribution performance on SST2, QQP, and MS-COCO -- datasets where potential spurious correlations are as yet uncharacterized. Human evaluation of ARGO groups shows that they contain well-defined, yet previously unstudied spurious correlations that lead to model errors.

Language models trained on massive prompted multitask datasets like T0 (Sanh et al., 2021) or FLAN (Wei et al., 2021a) can generalize to tasks unseen during training. We show that training on a carefully chosen subset of instances can outperform training on all available data on a variety of datasets. We assume access to a small number (250--1000) of unlabeled target task instances, select their nearest neighbors from a pool of multitask data, and use the retrieved data to train target task-specific models. Our method is more data-efficient than training a single multitask model, while still outperforming it by large margins. We evaluate across a diverse set of tasks not in the multitask pool we retrieve from, including those used to evaluate T0 and additional complex tasks including legal and scientific document QA. We retrieve small subsets of P3 (the collection of prompted datasets from which T0's training data was sampled) and finetune T5 models that outperform the 3-billion parameter variant of T0 (T0-3B) by 3--30% on 12 out of 14 evaluation datasets while using at most 2% of the data used to train T0-3B. These models also provide a better initialization than T0-3B for few-shot finetuning on target-task data, as shown by a 2--23% relative improvement over few-shot finetuned T0-3B models on 8 datasets. Our code is available at https://github.com/allenai/data-efficient-finetuning.

Chromosome analysis is essential for diagnosing genetic disorders. For hematologic malignancies, identification of somatic clonal aberrations by karyotype analysis remains the standard of care. However, karyotyping is costly and time-consuming because of the largely manual process and the expertise required in identifying and annotating aberrations. Efforts to automate karyotype analysis to date fell short in aberration detection. Using a training set of ~10k patient specimens and ~50k karyograms from over 5 years from the Fred Hutchinson Cancer Center, we created a labeled set of images representing individual chromosomes. These individual chromosomes were used to train and assess deep learning models for classifying the 24 human chromosomes and identifying chromosomal aberrations. The top-accuracy models utilized the recently introduced Topological Vision Transformers (TopViTs) with 2-level-block-Toeplitz masking, to incorporate structural inductive bias. TopViT outperformed CNN (Inception) models with >99.3% accuracy for chromosome identification, and exhibited accuracies >99% for aberration detection in most aberrations. Notably, we were able to show high-quality performance even in "few shot" learning scenarios. Incorporating the definition of clonality substantially improved both precision and recall (sensitivity). When applied to "zero shot" scenarios, the model captured aberrations without training, with perfect precision at >50% recall. Together these results show that modern deep learning models can approach expert-level performance for chromosome aberration detection. To our knowledge, this is the first study demonstrating the downstream effectiveness of TopViTs. These results open up exciting opportunities for not only expediting patient results but providing a scalable technology for early screening of low-abundance chromosomal lesions.

Predictive simulations of the shock-to-detonation transition (SDT) in heterogeneous energetic materials (EM) are vital to the design and control of their energy release and sensitivity. Due to the complexity of the thermo-mechanics of EM during the SDT, both macro-scale response and sub-grid mesoscale energy localization must be captured accurately. This work proposes an efficient and accurate multiscale framework for SDT simulations of EM. We employ deep learning to model the mesoscale energy localization of shock-initiated EM microstructures upon which prediction results are used to supply reaction progress rate information to the macroscale SDT simulation. The proposed multiscale modeling framework is divided into two stages. First, a physics-aware recurrent convolutional neural network (PARC) is used to model the mesoscale energy localization of shock-initiated heterogeneous EM microstructures. PARC is trained using direct numerical simulations (DNS) of hotspot ignition and growth within microstructures of pressed HMX material subjected to different input shock strengths. After training, PARC is employed to supply hotspot ignition and growth rates for macroscale SDT simulations. We show that PARC can play the role of a surrogate model in a multiscale simulation framework, while drastically reducing the computation cost and providing improved representations of the sub-grid physics. The proposed multiscale modeling approach will provide a new tool for material scientists in designing high-performance and safer energetic materials.