This paper presents a light-weight and accurate deep neural model for audiovisual emotion recognition. To design this model, the authors followed a philosophy of simplicity, drastically limiting the number of parameters to learn from the target datasets, always choosing the simplest earning methods: i) transfer learning and low-dimensional space embedding allows to reduce the dimensionality of the representations. ii) The isual temporal information is handled by a simple score-per-frame selection process, averaged across time. iii) A simple frame selection echanism is also proposed to weight the images of a sequence. iv) The fusion of the different modalities is performed at prediction level (late usion). We also highlight the inherent challenges of the AFEW dataset and the difficulty of model selection with as few as 383 validation equences. The proposed real-time emotion classifier achieved a state-of-the-art accuracy of 60.64 % on the test set of AFEW, and ranked 4th at he Emotion in the Wild 2018 challenge.

Purpose: To introduce a combined machine learning (ML) and physics-based image reconstruction framework that enables navigator-free, highly accelerated multishot echo planar imaging (msEPI), and demonstrate its application in high-resolution structural imaging. Methods: Singleshot EPI is an efficient encoding technique, but does not lend itself well to high-resolution imaging due to severe distortion artifacts and blurring. While msEPI can mitigate these artifacts, high-quality msEPI has been elusive because of phase mismatch arising from shot-to-shot physiological variations which disrupt the combination of the multiple-shot data into a single image. We employ Deep Learning to obtain an interim magnitude-valued image with minimal artifacts, which permits estimation of image phase variations due to shot-to-shot physiological changes. These variations are then included in a Joint Virtual Coil Sensitivity Encoding (JVC-SENSE) reconstruction to utilize data from all shots and improve upon the ML solution. Results: Our combined ML + physics approach enabled R=8-fold acceleration from 2 EPI-shots while providing 1.8-fold error reduction compared to the MUSSELS, a state-of-the-art reconstruction technique, which is also used as an input to our ML network. Using 3 shots allowed us to push the acceleration to R=10-fold, where we obtained a 1.7-fold error reduction over MUSSELS. Conclusion: Combination of ML and JVC-SENSE enabled navigator-free msEPI at higher accelerations than previously possible while using fewer shots, with reduced vulnerability to poor generalizability and poor acceptance of end-to-end ML approaches.

Many machine learning classifiers are vulnerable to adversarial attacks, inputs with perturbations designed to intentionally trigger misclassification. Modern adversarial methods either directly alter pixel colors, or "paint" colors onto a 3D shapes. We propose novel adversarial attacks that directly alter the geometry of 3D objects and/or manipulate the lighting in a virtual scene. We leverage a novel differentiable renderer that is efficient to evaluate and analytically differentiate. Our renderer generates images realistic enough for correct classification by common pre-trained models, and we use it to design physical adversarial examples that consistently fool these models. We conduct qualitative and quantitate experiments to validate our adversarial geometry and adversarial lighting attack capabilities.

This is a technical overview of the field of adversarial machine learning which has emerged to study vulnerabilities of machine learning approaches in adversarial settings and to develop techniques to make learning robust to adversarial manipulation. After reviewing machine learning concepts and approaches, as well as common use cases of these in adversarial settings, we present a general categorization of attacks on machine learning. We then address two major categories of attacks and associated defenses: decision-time attacks, in which an adversary changes the nature of instances seen by a learned model at the time of prediction in order to cause errors, and poisoning or training time attacks, in which the actual training dataset is maliciously modified. In our final chapter devoted to technical content, we discuss recent techniques for attacks on deep learning, as well as approaches for improving robustness of deep neural networks. We conclude with a discussion of several important issues in the area of adversarial learning that in our view warrant further research.

Embedded, continual learning for autonomous and adaptive behavior is a key application of neuromorphic hardware. However, neuromorphic implementations of embedded learning at large scales that are both flexible and efficient have been hindered by a lack of a suitable algorithmic framework. As a result, the most neuromorphic hardware is trained off-line on large clusters of dedicated processors or GPUs and transferred post hoc to the device. We address this by introducing the neural and synaptic array transceiver (NSAT), a neuromorphic computational framework facilitating flexible and efficient embedded learning by matching algorithmic requirements and neural and synaptic dynamics. NSAT supports event-driven supervised, unsupervised and reinforcement learning algorithms including deep learning. We demonstrate the NSAT in a wide range of tasks, including the simulation of Mihalas-Niebur neuron, dynamic neural fields, event-driven random back-propagation for event-based deep learning, event-based contrastive divergence for unsupervised learning, and voltage-based learning rules for sequence learning. We anticipate that this contribution will establish the foundation for a new generation of devices enabling adaptive mobile systems, wearable devices, and robots with data-driven autonomy.

In this paper, we propose a novel Question-Guided Hybrid Convolution (QGHC) network for Visual Question Answering (VQA). Most state-of-the-art VQA methods fuse the high-level textual and visual features from the neural network and abandon the visual spatial information when learning multi-modal features. To address these problems, question-guided kernels generated from the input question are designed to convolute with visual features for capturing the textual and visual relationship in the early stage. The question-guided convolution can tightly couple the textual and visual information but also introduce more parameters when learning kernels. We apply the group convolution, which consists of question-independent kernels and question-dependent kernels, to reduce the parameter size and alleviate over-fitting. The hybrid convolution can generate discriminative multi-modal features with fewer parameters. The proposed approach is also complementary to existing bilinear pooling fusion and attention based VQA methods. By integrating with them, our method could further boost the performance.

Technological development aims to produce generations of increasingly efficient robots able to perform complex tasks. This requires considerable efforts, from the scientific community, to find new algorithms that solve computer vision problems, such as object recognition. The diffusion of RGB-D cameras directed the study towards the research of new architectures able to exploit the RGB and Depth information. The project that is developed in this thesis concerns the realization of a new end-to-end architecture for the recognition of RGB-D objects called RCFusion. Our method generates compact and highly discriminative multi-modal features by combining complementary RGB and depth information representing different levels of abstraction. We evaluate our method on standard object recognition datasets, RGB-D Object Dataset and JHUIT-50. The experiments performed show that our method outperforms the existing approaches and establishes new state-of-the-art results for both datasets.

TensorFlow artificial intelligence framework is coming to Raspberry Pi. The developers were able to take advantage of the Tensor Flow in numerous ways to develop deep learning models and finally, the developers at Google were finally able to port the Tensorflow to Raspberry PI to develop the models on cheap hardware to perform things like image classification. Tensorflow AI can be used with Windows, Android and macOS but it would be difficult to find cheaper hardware than a Raspberry PI to run the ML framework. The latest release of the TensorFlow framework will be straightforward for the developers who are using Python as they can directly download the binaries from PIP from their Raspberry Pi and the users who are using the Raspberry Pi Foundation's Debian Stretch can directly installed the TensorFlow in their machines with two commands. Raspberry PI has total sales of 14 million units last year which explains the popularity of the developer board.

Artificial intelligence (AI) is going to make everything smarter, from the way we drive to how diseases are diagnosed. In the latest iteration of 101 Seconds, Qualcomm Technologies' head of Artificial Intelligence and Machine Learning Product Management, Gary Brotman, explains the difference between AI, machine learning, and deep learning, and the different ways on-device AI and AI-integrated machines will make our lives easier and more exciting.

Many of today's advanced scanning instruments have GPUs installed, which makes embedding AI algorithms or neural network architectures like deep learning onto the systems easier. Companies like GE Healthcare, Siemens Healthcare and Esaote use NVIDIA GPUs for visualization, helping accelerate scanned images by processing the magnetic resonance signals, ultrasound, or x-ray data streams.