Crowd counting problem aims to count the number of objects within an image or a frame in the videos and is usually solved by estimating the density map generated from the object location annotations. The values in the density map, by nature, take two possible states: zero indicating no object around, a non-zero value indicating the existence of objects and the value denoting the local object density. In contrast to traditional methods which do not differentiate the density prediction of these two states, we propose to use a dedicated network branch to predict the object/non-object mask and then combine its prediction with the input image to produce the density map. Our rationale is that the mask prediction could be better modeled as a binary segmentation problem and the difficulty of estimating the density could be reduced if the mask is known. A key to the proposed scheme is the strategy of incorporating the mask prediction into the density map estimator. To this end, we study five possible solutions, and via analysis and experimental validation we identify the most effective one. Through extensive experiments on five public datasets, we demonstrate the superior performance of the proposed approach over the baselines and show that our network could achieve the state-of-the-art performance.

Orthogonal frequency division multiplexing (OFDM) is one of the key technologies that are widely applied in current communication systems. Recently, artificial intelligence (AI)-aided OFDM receivers have been brought to the forefront to break the bottleneck of the traditional OFDM systems. In this paper, we investigate two AIaided OFDM receivers, data-driven fully connected-deep neural network (FC-DNN) receiver and model-driven ComNet receiver, respectively. We first study their performance under different channel models through simulation and then establish a real-time video transmission system using a 5G rapid prototyping (RaPro) system for over-the-air (OTA) test. To address the performance gap between the simulation and the OTA test caused by the discrepancy between the channel model for offline training and real environments, we develop a novel online training strategy, called SwitchNet receiver. The SwitchNet receiver is with a flexible and extendable architecture and can adapts to real channel by training one parameter online. The OTA test verifies its feasibility and robustness to real environments and indicates its potential for future communications systems. At the end of this paper, we discuss some challenges to inspire future research. P. Jiang, T. Wang, B. Han, X. Gao, J. Zhang and S. Jin are with the National Mobile Communications Research Laboratory, Southeast University, Nanjing 210096, China (email: wangtianqi@seu.edu.cn; C.-K. Wen is with the Institute of Communications Engineering, National Sun Yat-sen University, Kaohsiung 80424, Taiwan (email: ckwen@ieee.org). G. Y. Li is with the School of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA 30332 USA (email: liye@ece.gatech.edu). I. INTRODUCTION By introducing artificial intelligence (AI), intelligent communications can potentially address manychallenging issues in traditional communication systems. There have been many achievements in intelligent communications recently [1], [2], [3], including using AI for signal classification [4], multiple-input multiple-output (MIMO) detection [5], channel state information (CSI) feedback [6], [7], novel autoencoder-based end-to-end communication systems [8] and [9]. Orthogonal frequency division multiplexing (OFDM) has been proved to be an effective technique to deal with delay spread of wireless channels [10], [11]. OFDM receivers can be classified into two categories: linear and nonlinear receivers.

We propose a novel data augmentation method for labeled sentences called conditional BERT contextual augmentation. Data augmentation methods are often applied to prevent overfitting and improve generalization of deep neural network models. Recently proposed contextual augmentation augments labeled sentences by randomly replacing words with more varied substitutions predicted by language model. BERT demonstrates that a deep bidirectional language model is more powerful than either an unidirectional language model or the shallow concatenation of a forward and backward model. We retrofit BERT to conditional BERT by introducing a new conditional masked language model\footnote{The term "conditional masked language model" appeared once in original BERT paper, which indicates context-conditional, is equivalent to term "masked language model". In our paper, "conditional masked language model" indicates we apply extra label-conditional constraint to the "masked language model".} task. The well trained conditional BERT can be applied to enhance contextual augmentation. Experiments on six various different text classification tasks show that our method can be easily applied to both convolutional or recurrent neural networks classifier to obtain obvious improvement.

Spiking neural networks (SNNs) possess energy-efficient potential due to event-based computation. However, supervised training of SNNs remains a challenge as spike activities are non-differentiable. Previous SNNs training methods can basically be categorized into two classes, backpropagation-like training methods and plasticity-based learning methods. The former methods are dependent on energy-inefficient real-valued computation and non-local transmission, as also required in artificial neural networks (ANNs), while the latter either be considered biologically implausible or exhibit poor performance. Hence, biologically plausible (bio-plausible) high-performance supervised learning (SL) methods for SNNs remain deficient. In this paper, we proposed a novel bio-plausible SNN model for SL based on the symmetric spike-timing dependent plasticity (sym-STDP) rule found in neuroscience. By combining the sym-STDP rule with bio-plausible synaptic scaling and intrinsic plasticity of the dynamic threshold, our SNN model implemented SL well and achieved good performance in the benchmark recognition task (MNIST). To reveal the underlying mechanism of our SL model, we visualized both layer-based activities and synaptic weights using the t-distributed stochastic neighbor embedding (t-SNE) method after training and found that they were well clustered, thereby demonstrating excellent classification ability. As the learning rules were bio-plausible and based purely on local spike events, our model could be easily applied to neuromorphic hardware for online training and may be helpful for understanding SL information processing at the synaptic level in biological neural systems.

Deep convolutional neural networks have been widely used in numerous applications, but their demanding storage and computational resource requirements prevent their applications on mobile devices. Knowledge distillation aims to optimize a portable student network by taking the knowledge from a well-trained heavy teacher network. Traditional teacher-student based methods used to rely on additional fully-connected layers to bridge intermediate layers of teacher and student networks, which brings in a large number of auxiliary parameters. In contrast, this paper aims to propagate information from teacher to student without introducing new variables which need to be optimized. We regard the teacher-student paradigm from a new perspective of feature embedding. By introducing the locality preserving loss, the student network is encouraged to generate the low-dimensional features which could inherit intrinsic properties of their corresponding high-dimensional features from teacher network. The resulting portable network thus can naturally maintain the performance as that of the teacher network. Theoretical analysis is provided to justify the lower computation complexity of the proposed method. Experiments on benchmark datasets and well-trained networks suggest that the proposed algorithm is superior to state-of-the-art teacher-student learning methods in terms of computational and storage complexity.

We consider the Domain Adaptation problem, also known as the covariate shift problem, where the distributions that generate the training and test data differ while retaining the same labeling function. This problem occurs across a large range of practical applications, and is related to the more general challenge of transfer learning. Most recent work on the topic focuses on optimization techniques that are specific to an algorithm or practical use case rather than a more general approach. The sparse literature attempting to provide general bounds seems to suggest that efficient learning even under strong assumptions is not possible for covariate shift. Our main contribution is to recontextualize these results by showing that any Probably Approximately Correct (PAC) learnable concept class is still PAC learnable under covariate shift conditions with only a polynomial increase in the number of training samples. This approach essentially demonstrates that the Domain Adaptation learning problem is as hard as the underlying PAC learning problem, provided some conditions over the training and test distributions. We also present bounds for the rejection sampling algorithm, justifying it as a solution to the Domain Adaptation problem in certain scenarios.

Students are usually allocated tosupervisors for their projects by means of a centralized human decision maker or by means of interactions between students and staff members. The decision makers have to take into consideration the preferences of both students and supervisors with respect to the conduct of the project, as well as departmental constraintssuch as minimum and maximum levels of workload (in terms of supervision) for each supervisor. This situation results in an extremely time consuming process, and a suboptimal allocation due to a large and complex search space faced by human decision makers. Automating this process by applying artificial intelligence techniques may enhance the process in terms of satisfaction and performance of students with these individual projects. In this article, we present a genetic algorithm for matching students to supervisors accordingto both students' and supervisors' preferences and the constraints of the department. The rationale behind this problem is matching an appropriate student with a supervisor for the development of an individual project.The problem of matching students to supervisors, or students to projects [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13], is a subclass of the wider problem of matching between two sets, one of the most studied fields in computer sciencedue to its applications to a wide range of domains such as the hospital/residents (HR) or the college admission (CA) problem [14, 15, 16]. Particularly, the student-supervisor allocation problem solved in this article can be considered as an instance of the CA problem with lower and upper quotas, where the colleges are the supervisors, both colleges and students (i.e., supervisors andstudents in our case) have some representation of preferences on each other for the conduct of a project, and the minimum and maximum quotas are the minimum and maximum number of students to be supervised by staff members.

When I published the article "Understanding Blockchain" many of you wrote me to ask me if I could make one dedicated to Artificial Intelligence. The truth is that I hadn't had time to get on with it and before sharing anything, I wanted to finish some courses in order to add value to the recommendations. The problem with Artificial Intelligence is that it's much more fragmented, both technologically and in use cases, than Blockchain, making it a real challenge to condense all the information and share it meaningfully. Likewise, I have tried to make an effort in the summary of key concepts and in the compilation of interesting sources and resources, I hope it helps you as well as it did to me! Let's start with a little history. The timeline you see is taken from this article and it shows the most important milestones of Artificial Intelligence.

On March 18, 2018, at around 10 p.m., Elaine Herzberg was wheeling her bicycle across a street in Tempe, Arizona, when she was struck and killed by a self-driving car. Although there was a human operator behind the wheel, an autonomous system--artificial intelligence--was in full control. This incident, like others involving interactions between people and AI technologies, raises a host of ethical and proto-legal questions. What moral obligations did the system's programmers have to prevent their creation from taking a human life? And who was responsible for Herzberg's death? "Artificial intelligence" refers to systems that can be designed to take cues from their environment and, based on those inputs, proceed to solve problems, assess risks, make predictions, and take actions. In the era predating powerful computers and big data, such systems were programmed by humans and followed rules of human invention, but advances in technology have led to the development of new approaches.

As the success of deep learning reaches more grounds, one would like to also envision the potential limits of deep learning. This paper gives a first set of results proving that deep learning algorithms fail at learning certain efficiently learnable functions. Parity functions form the running example of our results and the paper puts forward a notion of low cross-predictability that defines a more general class of functions for which such failures tend to generalize (with examples in community detection and arithmetic learning). Recall that it is known that the class of neural networks (NNs) with polynomial network size can express any function that can be implemented in polynomial time, and that their sample complexity scales polynomially with the network size. The challenge is with the optimization error (the ERM is NP-hard), and the success behind deep learning is to train deep NNs with descent algorithms. The failures shown in this paper apply to training poly-size NNs on function distributions of low cross-predictability with a descent algorithm that is either run with limited memory per sample or that is initialized and run with enough randomness (exponentially small for GD). We further claim that such types of constraints are necessary to obtain failures, in that exact SGD with careful non-random initialization can learn parities. The cross-predictability notion has some similarity with the statistical dimension used in statistical query (SQ) algorithms, however the two definitions are different for reasons explained in the paper. The proof techniques are based on exhibiting algorithmic constraints that imply a statistical indistinguishability between the algorithm's output on the test model v.s.\ a null model, using information measures to bound the total variation distance.