We present a novel learning-based framework for face reenactment. The proposed method, known as ReenactGAN, is capable of transferring facial movements and expressions from monocular video input of an arbitrary person to a target person. Instead of performing a direct transfer in the pixel space, which could result in structural artifacts, we first map the source face onto a boundary latent space. A transformer is subsequently used to adapt the boundary of source face to the boundary of target face. Finally, a target-specific decoder is used to generate the reenacted target face. Thanks to the effective and reliable boundary-based transfer, our method can perform photo-realistic face reenactment. In addition, ReenactGAN is appealing in that the whole reenactment process is purely feed-forward, and thus the reenactment process can run in real-time (30 FPS on one GTX 1080 GPU). Dataset and model will be publicly available at https://wywu.github.io/projects/ReenactGAN/ReenactGAN.html

Convolutional neural networks have led to significant breakthroughs in the domain of medical image analysis. However, the task of breast cancer segmentation in whole-slide images (WSIs) is still underexplored. WSIs are large histopathological images with extremely high resolution. Constrained by the hardware and field of view, using high-magnification patches can slow down the inference process and using low-magnification patches can cause the loss of information. In this paper, we aim to achieve two seemingly conflicting goals for breast cancer segmentation: accurate and fast prediction. We propose a simple yet efficient framework Reinforced Auto-Zoom Net (RAZN) to tackle this task. Motivated by the zoom-in operation of a pathologist using a digital microscope, RAZN learns a policy network to decide whether zooming is required in a given region of interest. Because the zoom-in action is selective, RAZN is robust to unbalanced and noisy ground truth labels and can efficiently reduce overfitting. We evaluate our method on a public breast cancer dataset. RAZN outperforms both single-scale and multi-scale baseline approaches, achieving better accuracy at low inference cost.

Recent authors have proposed analyzing conditional reasoning through a notion of intervention on a simulation program, and have found a sound and complete axiomatization of the logic of conditionals in this setting. Here we extend this setting to the case of probabilistic simulation models. We give a natural definition of probability on formulas of the conditional language, allowing for the expression of counterfactuals, and prove foundational results about this definition. We also find an axiomatization for reasoning about linear inequalities involving probabilities in this setting. We prove soundness, completeness, and NP-completeness of the satisfiability problem for this logic.

Learning nonlinear dynamics from diffusion data is a challenging problem since the individuals observed may be different at different time points, generally following an aggregate behaviour. Existing work cannot handle the tasks well since they model such dynamics either directly on observations or enforce the availability of complete longitudinal individual-level trajectories. However, in most of the practical applications, these requirements are unrealistic: the evolving dynamics may be too complex to be modeled directly on observations, and individual-level trajectories may not be available due to technical limitations, experimental costs and/or privacy issues. To address these challenges, we formulate a model of diffusion dynamics as the {\em hidden stochastic process} via the introduction of hidden variables for flexibility, and learn the hidden dynamics directly on {\em aggregate observations} without any requirement for individual-level trajectories. We propose a dynamic generative model with Wasserstein distance for LEarninG dEep hidden Nonlinear Dynamics (LEGEND) and prove its theoretical guarantees as well. Experiments on a range of synthetic and real-world datasets illustrate that LEGEND has very strong performance compared to state-of-the-art baselines.

Research on automatic music generation has seen great progress due to the development of deep neural networks. However, the generation of multi-instrument music of arbitrary genres still remains a challenge. Existing research either works on lead sheets or multi-track piano-rolls found in MIDIs, but both musical notations have their limits. In this work, we propose a new task called lead sheet arrangement to avoid such limits. A new recurrent convolutional generative model for the task is proposed, along with three new symbolic-domain harmonic features to facilitate learning from unpaired lead sheets and MIDIs. Our model can generate lead sheets and their arrangements of eight-bar long. Audio samples of the generated result can be found at https://drive.google.com/open?id=1c0FfODTpudmLvuKBbc23VBCgQizY6-Rk

Original and learnt clauses in Conflict-Driven Clause Learning (CDCL) SAT solvers often contain redundant literals. This may have a negative impact on performance because redundant literals may deteriorate both the effectiveness of Boolean constraint propagation and the quality of subsequent learnt clauses. To overcome this drawback, we propose a clause vivification approach that eliminates redundant literals by applying unit propagation. The proposed clause vivification is activated before the SAT solver triggers some selected restarts, and only affects a subset of original and learnt clauses, which are considered to be more relevant according to metrics like the literal block distance (LBD). Moreover, we conducted an empirical investigation with instances coming from the hard combinatorial and application categories of recent SAT competitions. The results show that a remarkable number of additional instances are solved when the proposed approach is incorporated into five of the best performing CDCL SAT solvers (Glucose, TC_Glucose, COMiniSatPS, MapleCOMSPS and MapleCOMSPS_LRB). More importantly, the empirical investigation includes an in-depth analysis of the effectiveness of clause vivification. It is worth mentioning that one of the SAT solvers described here was ranked first in the main track of SAT Competition 2017 thanks to the incorporation of the proposed clause vivification. That solver was further improved in this paper and won the bronze medal in the main track of SAT Competition 2018.

New research proves that eyes might in fact be windows to the soul. Over the past few years, eye tracking technology has emerged as a field of much academic and corporate interest. With major acquisitions by Apple (SMI) and Oculus (The Eye Tribe), it's clear that major international companies regard eye tracking technology as an vital facet of Industry 4.0 -- particularly in its integration with virtual and augmented reality technologies (VR/AR), which involve persistent interaction with the human eye. In the study, researchers tracked 42 participants' eye movements while going about their day on a university campus, and matrixed these findings against user questionnaires. The results assert that machine learning can in fact deduce important personality traits with appropriate datasets -- with the algorithm reliably identifying four of the "Big Five" human personality traits: agreeableness, conscientiousness, extroversion, and neuroticism. In a statement from UniSA, Senior Lecturer of Psychology Dr. Tobias Loetscher explained that this research establishes a meaningful link between our eye motions and our innate and learned characteristics: People are always looking for improved, personalised [sic] services.

Nine days ago, we learned that Cambridge Analytica, the firm engaged by the Trump campaign to lead its digital strategy leading up to the 2016 United States presidential elections, illegitimately gained access to the Facebook data of more than 50 million users, many of them American voters. This revelation came on the heels of the announcement made last month by the Justice Department special counsel Robert Mueller of the indictment of 13 Russians who worked for the Internet Research Agency, a "troll farm" tied to the Kremlin, charging that they wielded fake social media accounts to influence the 2016 presidential elections. But as Facebook, Google, Twitter and like companies now contritely cover their tracks and comply with the government's requests, they simultaneously remain quiet about a critical trend that promises to subvert the nation's political integrity yet again if left unaddressed: the systemic integration of artificial intelligence into the same digital marketing technologies that were exploited by both Cambridge Analytica and the Internet Research Agency. According to the F.B.I.'s findings, the tactics used to date by Russia have, technologically speaking, not been particularly sophisticated. Those tactics have included the direct control of fake social media accounts and manual drafting of subversive messages. These were often timed for release with politically charged incidents in the real world -- including, for instance, the suicide bombings in Brussels, the declaration of Donald Trump as the Republican nominee and Mr. Trump's staging of a town hall in New Hampshire, each of which occurred weeks before election night in 2016.

President Donald Trump's July 19 executive order establishing the President's National Council for the American Worker is directed at preparing Americans for the workplace of the future. Although short on specifics, the order sends a powerful message about the need for revitalizing educational opportunities if Americans are to thrive in the era of big data, robots and artificial intelligence. The president's intent is to lay the groundwork for tackling a national "skills crisis." His order accepts that Americans need additional skills to fill the current 6.7 million job vacancies. In fact, the executive order gives official imprimatur to what many in industry and academia have feared for some time: "The economy is changing at a rapid pace because of the technology, automation, and artificial intelligence," and existing programs have "prepared Americans for the economy of the past."

Researchers have shown that they can use artificial intelligence (AI) to restore low-quality photos by exposing a neural network to only other low-quality photos, according to work presented at the International Conference on Machine Learning in Stockholm. The research was developed by representatives from Nvidia, a Santa Clara, California-based technology company, the Massachusetts Institute of Technology in Cambridge, Massachusetts, and Aalto University in Greater Helsinki, Finland. The team was able to clean up "grainy" photos by using deep learning to train the neural network with more than 50,000 example images, as explained in a news release from Nvidia. As the authors explained, their work has potential to be used in numerous industries, including radiology. "There are several real-world situations where obtaining clean training data is difficult: low-light photography (e.g., astronomical imaging), physically-based rendering, and magnetic resonance imaging," wrote author Jaakko Lehtinen, an associate professor at Aalto University, and colleagues.