随着社会越来越多的机器学习（ML）和深度学习（DL）进行各种智能解决方案，它将在普通开发实践和资源下规范ML / DL型号的ML / DL型号的常见措施变得越来越迫切。人们可以在共同的地面上基准和比较模型质量和性能。 MLCommons最近出现了来自工业和学术界的推动力，以协调这种努力。尽管其广泛采用作为标准化基准，但MLCommons推断只包括有限数量的ML / DL模型（实际上总共七种型号）。这显着限制了MLCommons推论的基准测试结果的一般性，因为研究界有更多的新颖ML / DL模型，解决了不同输入和输出方式的广泛问题。为了解决这种限制，我们提出了MLHARNESS，用于MLCommons推断的可扩展基准线束系统，具有三种独特的特点：（1）它根据MLCommons推理定义的标准基准过程，包括模型，数据集，DL框架和软件和硬件系统; （2）它为模型开发人员提供了一种简单且声明的方法，使其模型和数据集贡献到MLCommons推断; （3）它包括具有不同输入/输出模态的各种模型的支持，以便我们可以在不同的数据集，框架和硬件系统上缩放这些模型。该线束系统是在MLModelscope系统的顶部开发的，并将开放到社区。我们的实验结果表明，对于MLCommons推理基准测试，这种线束系统的卓越灵活性和可扩展性。

Gaze estimation is the fundamental basis for many visual tasks. Yet, the high cost of acquiring gaze datasets with 3D annotations hinders the optimization and application of gaze estimation models. In this work, we propose a novel Head-Eye redirection parametric model based on Neural Radiance Field, which allows dense gaze data generation with view consistency and accurate gaze direction. Moreover, our head-eye redirection parametric model can decouple the face and eyes for separate neural rendering, so it can achieve the purpose of separately controlling the attributes of the face, identity, illumination, and eye gaze direction. Thus diverse 3D-aware gaze datasets could be obtained by manipulating the latent code belonging to different face attributions in an unsupervised manner. Extensive experiments on several benchmarks demonstrate the effectiveness of our method in domain generalization and domain adaptation for gaze estimation tasks.

Three-dimensional (3D) ultrasound imaging technique has been applied for scoliosis assessment, but current assessment method only uses coronal projection image and cannot illustrate the 3D deformity and vertebra rotation. The vertebra detection is essential to reveal 3D spine information, but the detection task is challenging due to complex data and limited annotations. We propose VertMatch, a two-step framework to detect vertebral structures in 3D ultrasound volume by utilizing unlabeled data in semi-supervised manner. The first step is to detect the possible positions of structures on transverse slice globally, and then the local patches are cropped based on detected positions. The second step is to distinguish whether the patches contain real vertebral structures and screen the predicted positions from the first step. VertMatch develops three novel components for semi-supervised learning: for position detection in the first step, (1) anatomical prior is used to screen pseudo labels generated from confidence threshold method; (2) multi-slice consistency is used to utilize more unlabeled data by inputting multiple adjacent slices; (3) for patch identification in the second step, the categories are rebalanced in each batch to solve imbalance problem. Experimental results demonstrate that VertMatch can detect vertebra accurately in ultrasound volume and outperforms state-of-the-art methods. VertMatch is also validated in clinical application on forty ultrasound scans, and it can be a promising approach for 3D assessment of scoliosis.

Neural image classifiers are known to undergo severe performance degradation when exposed to input that exhibits covariate-shift with respect to the training distribution. Successful hand-crafted augmentation pipelines aim at either approximating the expected test domain conditions or to perturb the features that are specific to the training environment. The development of effective pipelines is typically cumbersome, and produce transformations whose impact on the classifier performance are hard to understand and control. In this paper, we show that recent Text-to-Image (T2I) generators' ability to simulate image interventions via natural-language prompts can be leveraged to train more robust models, offering a more interpretable and controllable alternative to traditional augmentation methods. We find that a variety of prompting mechanisms are effective for producing synthetic training data sufficient to achieve state-of-the-art performance in widely-adopted domain-generalization benchmarks and reduce classifiers' dependency on spurious features. Our work suggests that further progress in T2I generation and a tighter integration with other research fields may represent a significant step towards the development of more robust machine learning systems.

Despite recent progress towards scaling up multimodal vision-language models, these models are still known to struggle on compositional generalization benchmarks such as Winoground. We find that a critical component lacking from current vision-language models is relation-level alignment: the ability to match directional semantic relations in text (e.g., "mug in grass") with spatial relationships in the image (e.g., the position of the mug relative to the grass). To tackle this problem, we show that relation alignment can be enforced by encouraging the directed language attention from 'mug' to 'grass' (capturing the semantic relation 'in') to match the directed visual attention from the mug to the grass. Tokens and their corresponding objects are softly identified using the cross-modal attention. We prove that this notion of soft relation alignment is equivalent to enforcing congruence between vision and language attention matrices under a 'change of basis' provided by the cross-modal attention matrix. Intuitively, our approach projects visual attention into the language attention space to calculate its divergence from the actual language attention, and vice versa. We apply our Cross-modal Attention Congruence Regularization (CACR) loss to UNITER and improve on the state-of-the-art approach to Winoground.

During the deployment of deep neural networks (DNNs) on edge devices, many research efforts are devoted to the limited hardware resource. However, little attention is paid to the influence of dynamic power management. As edge devices typically only have a budget of energy with batteries (rather than almost unlimited energy support on servers or workstations), their dynamic power management often changes the execution frequency as in the widely-used dynamic voltage and frequency scaling (DVFS) technique. This leads to highly unstable inference speed performance, especially for computation-intensive DNN models, which can harm user experience and waste hardware resources. We firstly identify this problem and then propose All-in-One, a highly representative pruning framework to work with dynamic power management using DVFS. The framework can use only one set of model weights and soft masks (together with other auxiliary parameters of negligible storage) to represent multiple models of various pruning ratios. By re-configuring the model to the corresponding pruning ratio for a specific execution frequency (and voltage), we are able to achieve stable inference speed, i.e., keeping the difference in speed performance under various execution frequencies as small as possible. Our experiments demonstrate that our method not only achieves high accuracy for multiple models of different pruning ratios, but also reduces their variance of inference latency for various frequencies, with minimal memory consumption of only one model and one soft mask.

How to effectively leverage the plentiful existing datasets to train a robust and high-performance model is of great significance for many practical applications. However, a model trained on a naive merge of different datasets tends to obtain poor performance due to annotation conflicts and domain divergence.In this paper, we attempt to train a unified model that is expected to perform well across domains on several popularity segmentation datasets.We conduct a detailed analysis of the impact on model generalization from three aspects of data augmentation, training strategies, and model capacity.Based on the analysis, we propose a robust solution that is able to improve model generalization across domains.Our solution ranks 2nd on RVC 2022 semantic segmentation task, with a dataset only 1/3 size of the 1st model used.

Minimum Bayesian Risk Decoding (MBR) emerges as a promising decoding algorithm in Neural Machine Translation. However, MBR performs poorly with label smoothing, which is surprising as label smoothing provides decent improvement with beam search and improves generality in various tasks. In this work, we show that the issue arises from the un-consistency of label smoothing on the token-level and sequence-level distributions. We demonstrate that even though label smoothing only causes a slight change in the token-level, the sequence-level distribution is highly skewed. We coin the issue \emph{distributional over-smoothness}. To address this issue, we propose a simple and effective method, Distributional Cooling MBR (DC-MBR), which manipulates the entropy of output distributions by tuning down the Softmax temperature. We theoretically prove the equivalence between pre-tuning label smoothing factor and distributional cooling. Experiments on NMT benchmarks validate that distributional cooling improves MBR's efficiency and effectiveness in various settings.

Clustering analysis of sequence data continues to address many applications in engineering design, aided with the rapid growth of machine learning in applied science. This paper presents an unsupervised machine learning algorithm to extract defining characteristics of earthquake ground-motion records, also called latent features, to aid in ground-motion clustering and selection. In this context, a latent feature is a low dimensional machine-discovered spectral characteristic learned through nonlinear relationships of a neural network autoencoder. Clustering can be performed on the latent features and used to select a representative archetypal subgroup from a large ground-motion suite. The objective of efficient ground-motion selection is to choose records representative of what the structure will probabilistically experience in its lifetime. Three examples are presented to validate this approach, including a synthetic spectral dataset and spectra from field recorded ground-motion records. Deep embedding clustering of ground motion spectra improves on the results of static feature extraction, utilizing characteristics that represent the sparse spectral content of ground motions.

Accelerated MRI aims to find a pair of samplers and reconstructors to reduce acquisition time while maintaining the reconstruction quality. Most of the existing works focus on finding either sparse samplers with a fixed reconstructor or finding reconstructors with a fixed sampler. Recently, people have begun to consider learning samplers and reconstructors jointly. In this paper, we propose an alternating training framework for finding a good pair of samplers and reconstructors via deep reinforcement learning (RL). In particular, we propose a novel sparse-reward Partially Observed Markov Decision Process (POMDP) to formulate the MRI sampling trajectory. Compared to the existing works that utilize dense-reward POMDPs, the proposed sparse-reward POMDP is more computationally efficient and has a provable advantage over dense-reward POMDPs. We evaluate our method on fastMRI, a public benchmark MRI dataset, and it achieves state-of-the-art reconstruction performances.