This paper proposes a new method, that we call VisualBackProp, for visualizing which sets of pixels of the input image contribute most to the predictions made by the convolutional neural network (CNN). The method heavily hinges on exploring the intuition that the feature maps contain less and less irrelevant information to the prediction decision when moving deeper into the network. The technique we propose is dedicated for CNN-based systems for steering self-driving cars and is therefore required to run in real-time. This makes the proposed visualization method a valuable debugging tool which can be easily used during both training and inference. We justify our approach with theoretical arguments and confirm that the proposed method identifies sets of input pixels, rather than individual pixels, that collaboratively contribute to the prediction. We utilize the proposed visualization tool in the NVIDIA neural-network-based end-to-end learning system for autonomous driving, known as PilotNet. We demonstrate that VisualBackProp determines which elements in the road image most influence PilotNet's steering decision and indeed captures relevant objects on the road. The empirical evaluation furthermore shows the plausibility of the proposed approach on public road video data as well as in other applications and reveals that it compares favorably to the layer-wise relevance propagation approach, i.e. it obtains similar visualization results and achieves order of magnitude speed-ups.

We consider an efficient computational framework for speeding up several machine learning algorithms with almost no loss of accuracy. The proposed framework relies on projections via structured matrices that we call Structured Spinners, which are formed as products of three structured matrix-blocks that incorporate rotations. The approach is highly generic, i.e. i) structured matrices under consideration can either be fully-randomized or learned, ii) our structured family contains as special cases all previously considered structured schemes, iii) the setting extends to the non-linear case where the projections are followed by non-linear functions, and iv) the method finds numerous applications including kernel approximations via random feature maps, dimensionality reduction algorithms,new fast cross-polytope LSH techniques, deep learning, convex optimization algorithms via Newton sketches, quantization with random projection trees, and more. The proposed framework comes with theoretical guarantees characterizing the capacity of the structured model in reference to its unstructured counterpart and is based on a general theoretical principle that we describe in the paper. As a consequence of our theoretical analysis, we provide the first theoretical guarantees for one of the most efficient existing LSH algorithms based on the HD 3 HD 2 HD 1 structured matrix [Andoni et al., 2015]. The exhaustive experimental evaluation confirms the accuracy and efficiency of structured spinners for a variety of different applications.

We consider the hashing mechanism for constructing binary embeddings, that involves pseudo-random projections followed by nonlinear (sign function) mappings. The pseudo-random projection is described by a matrix, where not all entries are independent random variables but instead a fixed “budget of randomness” is distributed across the matrix. Such matrices can be edfficiently stored in sub-quadratic or even linear space, provide reduction in randomness usage (i.e. number of required random values), and very often lead to computational speed ups. We prove several theoretical results showing that projections via various structured matrices followed by nonlinear mappings accurately preserve the angular distance between input high-dimensional vectors. To the best of our knowledge, these results are the first that give theoretical ground for the use of general structured matrices in the nonlinear setting. We empirically verify our theoretical findings and show the dependence of learning via structured hashed projections on the performance of neural network as well as nearest neighbor classifier.

In this study, a multi-level type multi-phase resonant converter is presented for high power wireless EV charging applications. As an alternative to the traditional frequency and phase shift control methods, a hybrid phase-frequency control strategy is implemented to improve the system efficiency. In order to confirm the proposed converter and control technique, a laboratory prototype wireless EV charger is designed using 8 inches air gap coreless transformer and rectifier. The proposed control is compared with the conventional control methods for various load conditions at the different power levels. The experimental results show that the proposed converter is within the desired frequency range while regulating output from 0 to 15 kW with 750 V input DC bus voltage.

In this study, a CLL resonant converter with different control strategies; frequency, dc-link, and phase shift control are compared and a hybrid control is proposed to increase the efficiency of the system for Electric Vehicle (EV) battery charger applications. The comparison analysis is examined considering the battery charging control algorithms constant current (CC) and constant voltage (CV) with modeling the battery as a variable load. In order to improve charging efficiency, the system should operate in a close to resonant frequency, achieving ZVS and reducing circulating current. The presented control is achieved by staying around the resonant frequency in all charging conditions. The designed dc/dc resonant converter, fed by a power-factor-corrector (PFC) with 80-120 V output, provides 120 V / 10 A at 1.2 kW maximum power. The application results show that the proposed hybrid control improves overall efficiency 2 %.

A new family of Class D resonant inverters is proposed in this paper. Multiple identical series resonant inverters are paralleled using intercell transformers to form phase-controlled multiphase resonant inverter with a common resonant circuit. Inverters can operate at constant frequency utilizing phase-shift control to regulate output. A frequency-domain analysis of the proposed family is performed. An experimental prototype of a three-phase resonant inverter with common resonant circuit was built and extensively tested at an output power of 550 W and switching frequency of 167 kHz.