We propose a novel scene flow estimation approach to capture and infer 3D motions from point clouds. Estimating 3D motions for point clouds is challenging, since a point cloud is unordered and its density is significantly non-uniform. Such unstructured data poses difficulties in matching corresponding points between point clouds, leading to inaccurate flow estimation. We propose a novel architecture named Sparse Convolution-Transformer Network (SCTN) that equips the sparse convolution with the transformer. Specifically, by leveraging the sparse convolution, SCTN transfers irregular point cloud into locally consistent flow features for estimating continuous and consistent motions within an object/local object part. We further propose to explicitly learn point relations using a point transformer module, different from exiting methods. We show that the learned relation-based contextual information is rich and helpful for matching corresponding points, benefiting scene flow estimation. In addition, a novel loss function is proposed to adaptively encourage flow consistency according to feature similarity. Extensive experiments demonstrate that our proposed approach achieves a new state of the art in scene flow estimation. Our approach achieves an error of 0.038 and 0.037 (EPE3D) on FlyingThings3D and KITTI Scene Flow respectively, which significantly outperforms previous methods by large margins.

Quantization techniques applied to the inference of deep neural networks have enabled fast and efficient execution on resource-constraint devices. The success of quantization during inference has motivated the academic community to explore fully quantized training, i.e. quantizing back-propagation as well. However, effective gradient quantization is still an open problem. Gradients are unbounded and their distribution changes significantly during training, which leads to the need for dynamic quantization. As we show, dynamic quantization can lead to significant memory overhead and additional data traffic slowing down training. We propose a simple alternative to dynamic quantization, in-hindsight range estimation, that uses the quantization ranges estimated on previous iterations to quantize the present. Our approach enables fast static quantization of gradients and activations while requiring only minimal hardware support from the neural network accelerator to keep track of output statistics in an online fashion. It is intended as a drop-in replacement for estimating quantization ranges and can be used in conjunction with other advances in quantized training. We compare our method to existing methods for range estimation from the quantized training literature and demonstrate its effectiveness with a range of architectures, including MobileNetV2, on image classification benchmarks (Tiny ImageNet & ImageNet).

Human pose estimation is a major computer vision problem with applications ranging from augmented reality and video capture to surveillance and movement tracking. In the medical context, the latter may be an important biomarker for neurological impairments in infants. Whilst many methods exist, their application has been limited by the need for well annotated large datasets and the inability to generalize to humans of different shapes and body compositions, e.g. children and infants. In this paper we present a novel method for learning pose estimators for human adults and infants in an unsupervised fashion. We approach this as a learnable template matching problem facilitated by deep feature extractors. Human-interpretable landmarks are estimated by transforming a template consisting of predefined body parts that are characterized by 2D Gaussian distributions. Enforcing a connectivity prior guides our model to meaningful human shape representations. We demonstrate the effectiveness of our approach on two different datasets including adults and infants.

The performance of deep reinforcement learning methods prone to degenerate when applied to environments with non-stationary dynamics. In this paper, we utilize the latent context recurrent encoders motivated by recent Meta-RL materials, and propose the Latent Context-based Soft Actor Critic (LC-SAC) method to address aforementioned issues. By minimizing the contrastive prediction loss function, the learned context variables capture the information of the environment dynamics and the recent behavior of the agent. Then combined with the soft policy iteration paradigm, the LC-SAC method alternates between soft policy evaluation and soft policy improvement until it converges to the optimal policy. Experimental results show that the performance of LC-SAC is significantly better than the SAC algorithm on the MetaWorld ML1 tasks whose dynamics changes drasticly among different episodes, and is comparable to SAC on the continuous control benchmark task MuJoCo whose dynamics changes slowly or doesn't change between different episodes. In addition, we also conduct relevant experiments to determine the impact of different hyperparameter settings on the performance of the LC-SAC algorithm and give the reasonable suggestions of hyperparameter setting.

Deep neural networks that yield human interpretable decisions by architectural design have lately become an increasingly popular alternative to post hoc interpretation of traditional black-box models. Among these networks, the arguably most widespread approach is so-called prototype learning, where similarities to learned latent prototypes serve as the basis of classifying an unseen data point. In this work, we point to an important shortcoming of such approaches. Namely, there is a semantic gap between similarity in latent space and similarity in input space, which can corrupt interpretability. We design two experiments that exemplify this issue on the so-called ProtoPNet. Specifically, we find that this network's interpretability mechanism can be led astray by intentionally crafted or even JPEG compression artefacts, which can produce incomprehensible decisions. We argue that practitioners ought to have this shortcoming in mind when deploying prototype-based models in practice.

We present PSEUDo, an adaptive feature learning technique for exploring visual patterns in multi-track sequential data. Our approach is designed with the primary focus to overcome the uneconomic retraining requirements and inflexible representation learning in current deep learning-based systems. Multi-track time series data are generated on an unprecedented scale due to increased sensors and data storage. These datasets hold valuable patterns, like in neuromarketing, where researchers try to link patterns in multivariate sequential data from physiological sensors to the purchase behavior of products and services. But a lack of ground truth and high variance make automatic pattern detection unreliable. Our advancements are based on a novel query-aware locality-sensitive hashing technique to create a feature-based representation of multivariate time series windows. Most importantly, our algorithm features sub-linear training and inference time. We can even accomplish both the modeling and comparison of 10,000 different 64-track time series, each with 100 time steps (a typical EEG dataset) under 0.8 seconds. This performance gain allows for a rapid relevance feedback-driven adaption of the underlying pattern similarity model and enables the user to modify the speed-vs-accuracy trade-off gradually. We demonstrate superiority of PSEUDo in terms of efficiency, accuracy, and steerability through a quantitative performance comparison and a qualitative visual quality comparison to the state-of-the-art algorithms in the field. Moreover, we showcase the usability of PSEUDo through a case study demonstrating our visual pattern retrieval concepts in a large meteorological dataset. We find that our adaptive models can accurately capture the user's notion of similarity and allow for an understandable exploratory visual pattern retrieval in large multivariate time series datasets.

In recent years, data and computing resources are typically distributed in the devices of end users, various regions or organizations. Because of laws or regulations, the distributed data and computing resources cannot be directly shared among different regions or organizations for machine learning tasks. Federated learning emerges as an efficient approach to exploit distributed data and computing resources, so as to collaboratively train machine learning models, while obeying the laws and regulations and ensuring data security and data privacy. In this paper, we provide a comprehensive survey of existing works for federated learning. We propose a functional architecture of federated learning systems and a taxonomy of related techniques. Furthermore, we present the distributed training, data communication, and security of FL systems. Finally, we analyze their limitations and propose future research directions.

Conversational agent (CA) systems have been applied to healthcare domain, but there is no such a system to answer consumers regarding DS use, although widespread use of DS. In this study, we develop the first CA system for DS use. Methods: Our CA system for DS use developed on the MindeMeld framework, consists of three components: question understanding, DS knowledge base, and answer generation. We collected and annotated 1509 questions to develop natural language understanding module (e.g., question type classifier, named entity recognizer) which was then integrated into MindMeld framework. CA then queries the DS knowledge base (i.e., iDISK) and generates answers using rule-based slot filling techniques. We evaluated algorithms of each component and the CA system as a whole. Results: CNN is the best question classifier with F1 score of 0.81, and CRF is the best named entity recognizer with F1 score of 0.87. The system achieves an overall accuracy of 81% and an average score of 1.82 with succ@3 score as 76.2% and succ@2 as 66% approximately. Conclusion: This study develops the first CA system for DS use using MindMeld framework and iDISK domain knowledge base.

Austrian company Tec-Innovation recently unveiled smart shoes that use ultrasonic sensors to help people suffering from blindness of vision impairment to detect obstacles up to four meters away. Known as InnoMake, the smart shoe aims to become a modern alternative to the decades-old walking stick that millions of people around the world depend on to get around as safely as possible. The currently available model relies on sensors to detect obstacles and warns the wearer via vibration and an audible alert sounded on a Bluetooth-linked smartphone. That sounds impressive enough, but the company is already working on a much more advanced version that incorporates cameras and artificial intelligence to not only detect obstacles but also their nature. Tec-Innovation partnered with Austria's Graz University of Technology to develop of state-of-the-art deep-learning algorithms modeled on neural networks that can analyze the information provided by sensors and cameras incorporated in the InnoMake shoe to determine whether an area is free obstacles and safe to walk on, and also distinguish between various types of obstacles. "Not only is the warning that I am facing an obstacle relevant, but also the information about what kind of obstacle I am facing.

If you have built Deep Neural Networks before, you might know that it can involve a lot of experimentation. In this article, I will share with you some useful tips and guidelines that you can use to better build better deep learning models. These tricks should make it a lot easier for you to develop a good network. You can pick and choose which tips you use, as some will be more helpful for the projects you are working on. Not everything mentioned in this article will straight up improve your models' performance.