We reside in a world, where we are constantly surrounded by deep learning algorithms be it for the good or worse cause. From the Netflix recommendation system to Tesla's autonomous cars, Deep Learning is leaving its stark appearance in our lives. This quaint & lucid technology is something that can impeach the whole thousand years old human civilization with just 4–5 years of training. You've probably suggested this article because Deep Learning thinks you should see it. Now let's jump right in! Deep learning is an extended arm of machine learning algorithms that teaches computers to do what is inherited naturally to humans i.e. learn by examples.

This article is part of Demystifying AI, a series of posts that (try to) disambiguate the jargon and myths surrounding AI. To human observers, the following two images are identical. But researchers at Google showed in 2015 that a popular object detection algorithm classified the left image as "panda" and the right one as "gibbon." And oddly enough, it had more confidence in the gibbon image. The algorithm in question was GoogLeNet, a convolutional neural network architecture that won the 2014 ImageNet Large Scale Visual Recognition Challenge (ILSVRC 2014). The right image is an "adversarial example."

In the continuing theme of higher level tools to improve developing useful applications, today we'll visit feature engineering in a changing environment. Artificial intelligence (AI) is increasingly used to analyze data, and deep learning (DL) is one of the more complex aspects of AI. In multiple forums, I've discussed the need to move past heavy reliance on not just pure coding, but even past the basic frameworks discussed by DL programmers. One of the keys to the complexity is figuring out the right data attributes, or features, which matter to any system. As tricky as that is the first time, it needs to be a repeatable process, as environments change, and systems must change with them. Defining the initial feature set is important, but it's not the end of the game.

Spiking Neural Networks (SNNs) are gaining interest due to their event-driven processing which potentially consumes low power/energy computations in hardware platforms, while offering unsupervised learning capability due to the spike-timing-dependent plasticity (STDP) rule. However, state-of-the-art SNNs require a large memory footprint to achieve high accuracy, thereby making them difficult to be deployed on embedded systems, for instance on battery-powered mobile devices and IoT Edge nodes. Towards this, we propose FSpiNN, an optimization framework for obtaining memory- and energy-efficient SNNs for training and inference processing, with unsupervised learning capability while maintaining accuracy. It is achieved by (1) reducing the computational requirements of neuronal and STDP operations, (2) improving the accuracy of STDP-based learning, (3) compressing the SNN through a fixed-point quantization, and (4) incorporating the memory and energy requirements in the optimization process. FSpiNN reduces the computational requirements by reducing the number of neuronal operations, the STDP-based synaptic weight updates, and the STDP complexity. To improve the accuracy of learning, FSpiNN employs timestep-based synaptic weight updates, and adaptively determines the STDP potentiation factor and the effective inhibition strength. The experimental results show that, as compared to the state-of-the-art work, FSpiNN achieves 7.5x memory saving, and improves the energy-efficiency by 3.5x on average for training and by 1.8x on average for inference, across MNIST and Fashion MNIST datasets, with no accuracy loss for a network with 4900 excitatory neurons, thereby enabling energy-efficient SNNs for edge devices/embedded systems.

To get clear street-view and photo-realistic simulation in autonomous driving, we present an automatic video inpainting algorithm that can remove traffic agents from videos and synthesize missing regions with the guidance of depth/point cloud. By building a dense 3D map from stitched point clouds, frames within a video are geometrically correlated via this common 3D map. In order to fill a target inpainting area in a frame, it is straightforward to transform pixels from other frames into the current one with correct occlusion. Furthermore, we are able to fuse multiple videos through 3D point cloud registration, making it possible to inpaint a target video with multiple source videos. The motivation is to solve the long-time occlusion problem where an occluded area has never been visible in the entire video. To our knowledge, we are the first to fuse multiple videos for video inpainting. To verify the effectiveness of our approach, we build a large inpainting dataset in the real urban road environment with synchronized images and Lidar data including many challenge scenes, e.g., long time occlusion. The experimental results show that the proposed approach outperforms the state-of-the-art approaches for all the criteria, especially the RMSE (Root Mean Squared Error) has been reduced by about 13%.

Due to the characteristics of COVID-19, the epidemic develops rapidly and overwhelms health service systems worldwide. Many patients suffer from systemic life-threatening problems and need to be carefully monitored in ICUs. Thus the intelligent prognosis is in an urgent need to assist physicians to take an early intervention, prevent the adverse outcome, and optimize the medical resource allocation. However, in the early stage of the epidemic outbreak, the data available for analysis is limited due to the lack of effective diagnostic mechanisms, rarity of the cases, and privacy concerns. In this paper, we propose a deep-learning-based approach, CovidCare, which leverages the existing electronic medical records to enhance the prognosis for inpatients with emerging infectious diseases. It learns to embed the COVID-19-related medical features based on massive existing EMR data via transfer learning. The transferred parameters are further trained to imitate the teacher model's representation behavior based on knowledge distillation, which embeds the health status more comprehensively in the source dataset. We conduct the length of stay prediction experiments for patients on a real-world COVID-19 dataset. The experiment results indicate that our proposed model consistently outperforms the comparative baseline methods. CovidCare also reveals that, 1) hs-cTnI, hs-CRP and Platelet Counts are the most fatal biomarkers, whose abnormal values usually indicate emergency adverse outcome. 2) Normal values of gamma-GT, AP and eGFR indicate the overall improvement of health. The medical findings extracted by CovidCare are empirically confirmed by human experts and medical literatures.

Tracking a car or a person in a city is crucial for urban safety management. How can we complete the task with minimal number of spatiotemporal searches from massive camera records? This paper proposes a strategy named IHMs (Intermediate Searching at Heuristic Moments): each step we figure out which moment is the best to search according to a heuristic indicator, then at that moment search locations one by one in descending order of predicted appearing probabilities, until a search hits; iterate this step until we get the object's current location. Five searching strategies are compared in experiments, and IHMs is validated to be most efficient, which can save up to 1/3 total costs. This result provides an evidence that "searching at intermediate moments can save cost".

Many social services including online dating, social media, recruitment and online learning, largely rely on \matching people with the right people". The success of these services and the user experience with them often depends on their ability to match users. Reciprocal Recommender Systems (RRS) arose to facilitate this process by identifying users who are a potential match for each other, based on information provided by them. These systems are inherently more complex than user-item recommendation approaches and unidirectional user recommendation services, since they need to take into account both users' preferences towards each other in the recommendation process. This entails not only predicting accurate preference estimates as classical recommenders do, but also defining adequate fusion processes for aggregating user-to-user preferential information. The latter is a crucial and distinctive, yet barely investigated aspect in RRS research. This paper presents a snapshot analysis of the extant literature to summarize the state-of-the-art RRS research to date, focusing on the fundamental features that differentiate RRSs from other classes of recommender systems. Following this, we discuss the challenges and opportunities for future research on RRSs, with special focus on (i) fusion strategies to account for reciprocity and (ii) emerging application domains related to social recommendation.

The network calculus (NC) analysis takes a simple model consisting of a network of schedulers and data flows crossing them. A number of analysis "building blocks" can then be applied to capture the model without imposing pessimistic assumptions like self-contention on tandems of servers. Yet, adding pessimism cannot always be avoided. To compute the best bound on a single flow's end-to-end delay thus boils down to finding the least pessimistic contention models for all tandems of schedulers in the network - and an exhaustive search can easily become a very resource intensive task. The literature proposes a promising solution to this dilemma: a heuristic making use of machine learning (ML) predictions inside the NC analysis. While results of this work were promising in terms of delay bound quality and computational effort, there is little to no insight on when a prediction is made or if the trained algorithm can achieve similarly striking results in networks vastly differing from its training data. In this paper, we address these pending questions. We evaluate the influence of the training data and its features on accuracy, impact and scalability. Additionally, we contribute an extension of the method by predicting the best $n$ contention model alternatives in order to achieve increased robustness for its application outside the training data. Our numerical evaluation shows that good accuracy can still be achieved on large networks although we restrict the training to networks that are two orders of magnitude smaller.

Deep learning based models are relatively large, and it is hard to deploy such models on resource-limited devices such as mobile phones and embedded devices. One possible solution is knowledge distillation whereby a smaller model (student model) is trained by utilizing the information from a larger model (teacher model). In this paper, we present a survey of knowledge distillation techniques applied to deep learning models. To compare the performances of different techniques, we propose a new metric called distillation metric. Distillation metric compares different knowledge distillation algorithms based on sizes and accuracy scores. Based on the survey, some interesting conclusions are drawn and presented in this paper.