--Recent state-of-the-art video generation systems employ Generative Adversarial Networks (GANs) or V ariational Autoencoders (V AEs) to produce novel videos. However, V AE models typically produce blurry outputs when faced with sub-optimal conditioning of the input, and GANs are known to be unstable for large output sizes. In addition, the output videos of these models are difficult to evaluate, partly because the GAN loss function is not an accurate measure of convergence. In this work, we propose using a state-of-the-art neural flow generator called Glow to generate videos conditioned on a textual label, one frame at a time. Neural flow models are more stable than standard GANs, as they only optimize a single cross entropy loss function, which is monotonic and avoids the circular convergence issues of the GAN minimax objective. In addition, we also show how to condition Glow on external context, while still preserving the invertible nature of each "flow" layer . Finally, we evaluate the proposed Glow model by calculating cross entropy on a held-out validation set of videos, in order to compare multiple versions of the proposed model via an ablation study. We show generated videos and discuss future improvements. I NTRODUCTION Text-to-video generation is the process by which a model conditions on text, and produces a video based on that text description. This is the exact opposite of video captioning, which aims to produce a caption that would describe a given video [6]. It may be argued that text-to-video is a harder task, as there are many more degrees of freedom in pixel space.

Adaptive and Iteratively Improving Recurrent Lateral ConnectionsBarak Battash Lior Wolf Tel Aviv University Facebook AI Research & Tel Aviv University Abstract The current leading computer vision models are typically feed forward neural models, in which the output of one computational block is passed to the next one sequentially. This is in sharp contrast to the organization of the primate visual cortex, in which feedback and lateral connections are abundant. In this work, we propose a computational model for the role of lateral connections in a given block, in which the weights of the block vary dynamically as a function of its activations, and the input from the upstream blocks is iteratively reintroduced. We demonstrate how this novel architectural modification can lead to sizable gains in performance, when applied to visual action recognition without pretraining and that it outperforms the literature architectures with recurrent feedback processing on ImageNet. 1 Introduction Rapid exposure experiments in primates teach us that image recognition occurs as early as the first 100 msec of visual perception, a time budget that suffices only for feed-forward inference, due to the relatively slow nature of biological neurons (Perrett and Oram, 1993; Thorpe et al., 1996). However, anatomical studies have shown that feedback connections are prevalent in the cortex (Douglas and Martin, 2004; Felleman and Essen, 1991). As one striking example, the feedforward input from LGN to V1 in cats constitutes only five percent of the total input to V1, the rest being lateral and feedback connections (Binzegger et al., 2004). In fact, lateral connections, which are projections from a layer to itself, are even more prevalent than feedback connections that project from downstream layers upstream. One possible conjecture would be that feedback (including lateral) connections play roles that are replaced by other mechanisms in the current deep learning literature. For example, they could play a role in training the biological neural network, or they can form attention mechanisms, which are captured by attention (Sermanet et al., 2014) and self-attention (Parikh et al., 2016) blocks in modern neural networks. Similarly, one can claim that such connections are required due to the limitations of the biological computational, but may not be necessary in artificial neural networks, which can be extremely deep (Liao and Poggio, 2016).

In this work we present an adversarial training algorithm that exploits correlations in video to learn --without supervision-- an image generator model with a disentangled latent space. The proposed methodology requires only a few modifications to the standard algorithm of Generative Adversarial Networks (GAN) and involves training with sets of frames taken from short videos. We train our model over two datasets of face-centered videos which present different people speaking or moving the head: VidTIMIT and YouTube Faces datasets. We found that our proposal allows us to split the generator latent space into two subspaces. One of them controls content attributes, those that do not change along short video sequences. For the considered datasets, this is the identity of the generated face. The other subspace controls motion attributes, those attributes that are observed to change along short videos. We observed that these motion attributes are face expressions, head orientation, lips and eyes movement. The presented experiments provide quantitative and qualitative evidence supporting that the proposed methodology induces a disentangling of this two kinds of attributes in the latent space.

The Transformer architecture has become increasingly popular over the past couple of years, owing to its impressive performance on a number of natural language processing (NLP) tasks. However, it may be argued that the Transformer architecture lacks an explicit hierarchical representation, as all computations occur on word-level representations alone, and therefore, learning structure poses a challenge for Transformer models. In the present work, we introduce hierarchical processing into the Transformer model, taking inspiration from the U-Net architecture, popular in computer vision for its hierarchical view of natural images. We propose a novel architecture that combines ideas from Transformer and U-Net models to incorporate hierarchy at multiple levels of abstraction. We empirically demonstrate that the proposed architecture outperforms the vanilla Transformer and strong baselines in the chit-chat dialogue and machine translation domains.

Recent dialogue approaches operate by reading each word in a conversation history, and aggregating accrued dialogue information into a single state. This fixed-size vector is not expandable and must maintain a consistent format over time. Other recent approaches exploit an attention mechanism to extract useful information from past conversational utterances, but this introduces an increased computational complexity. In this work, we explore the use of the Neural Turing Machine (NTM) to provide a more permanent and flexible storage mechanism for maintaining dialogue coherence. Specifically, we introduce two separate dialogue architectures based on this NTM design. The first design features a sequence-to-sequence architecture with two separate NTM modules, one for each participant in the conversation. The second memory architecture incorporates a single NTM module, which stores parallel context information for both speakers. This second design also replaces the sequence-to-sequence architecture with a neural language model, to allow for longer context of the NTM and greater understanding of the dialogue history. We report perplexity performance for both models, and compare them to existing baselines.

A BSTRACT State-of-the-art neural machine translation methods employ massive amounts of parameters. Drastically reducing computational costs of such methods without affecting performance has been up to this point unsolved. In this work, we propose a quantization strategy tailored to the Transformer (V aswani et al., 2017) architecture. We evaluate our method on the WMT14 EN-FR and WMT14 EN-DE translation tasks and achieve state-of-the-art quantization results for the Transformer, obtaining no loss in BLEU scores compared to the non-quantized baseline. We further compress the Transformer by showing that, once the model is trained, a good portion of the nodes in the encoder can be removed without causing any loss in BLEU. 1 I NTRODUCTION Neural machine translation methods have achieved impressive results lately (Ahmed et al., 2017; Ott et al., 2018; Edunov et al., 2018). Having been proposed only recently (Kalchbrenner & Blunsom, 2013; Sutskever et al., 2014; Cho et al., 2014), many great work have led the field to move forward quickly. Bahdanau et al. (2014) introduced an attention mechanism, allowing the decoder to attend to any hidden state generated by the encoder. Multiple improvements to their approach have been proposed, such as multiplicative attention (Luong et al., 2015) and more recently multi-head self-attention (V aswani et al., 2017).

Automated classification of metadata of research data by their discipline(s) of research can be used in scientometric research, by repository service providers, and in the context of research data aggregation services. Openly available metadata of the DataCite index for research data were used to compile a large training and evaluation set comprised of 609,524 records, which is published alongside this paper. These data allow to reproducibly assess classification approaches, such as tree-based models and neural networks. According to our experiments with 20 base classes (multi-label classification), multi-layer perceptron models perform best with a f1-macro score of 0.760 closely followed by Long Short-Term Memory models (f1-macro score of 0.755). A possible application of the trained classification models is the quantitative analysis of trends towards interdisciplinarity of digital scholarly output or the characterization of growth patterns of research data, stratified by discipline of research. Both applications perform at scale with the proposed models which are available for re-use.

A detailed understanding of users contributes to the understanding of the Web's evolution, and to the development of Web applications. Although for new Web platforms such a study is especially important, it is often jeopar dized by the lack of knowledge about novel phenomena due to the sparsity of data. Akin to human transfer of experiences from one domain to the next, transfer learning as a subfield of machine learning adapts knowledge acquired in one domain to a new domain . We systematically investigate how the concept of transfer learning may be applied to the study of users on newly created (emerging) Web platforms, and propose our transfer learning - based approach, TraNet. We show two use cases where TraNet is applied to tasks involving the identification of user trust and roles on different Web platforms. We compare the performance of TraNet with other approaches and find that our approach can best transfer knowledge on users across platforms in the given tasks.

In this paper, we propose a gamification approach as a novel framework for smart building infrastructure with the goal of motivating human occupants to reconsider personal energy usage and to have positive effects on their environment. Human interaction in the context of cyber-physical systems is a core component and consideration in the implementation of any smart building technology. Research has shown that the adoption of human-centric building services and amenities leads to improvements in the operational efficiency of these cyber-physical systems directed towards controlling building energy usage. We introduce a strategy in form of a game-theoretic framework that incorporates humans-in-the-loop modeling by creating an interface to allow building managers to interact with occupants and potentially incentivize energy efficient behavior. Prior works on game theoretic analysis typically rely on the assumption that the utility function of each individual agent is known a priori. Instead, we propose novel utility learning framework for benchmarking that employs robust estimations of occupant actions towards energy efficiency. To improve forecasting performance, we extend the utility learning scheme by leveraging deep bi-directional recurrent neural networks. Using the proposed methods on data gathered from occupant actions for resources such as room lighting, we forecast patterns of energy resource usage to demonstrate the prediction performance of the methods. The results of our study show that we can achieve a highly accurate representation of the ground truth for occupant energy resource usage. We also demonstrate the explainable nature on human decision making towards energy usage inherent in the dataset using graphical lasso and granger causality algorithms. Finally, we open source the de-identified, high-dimensional data pertaining to the energy game-theoretic framework.

Reservoir computers (RCs) and recurrent neural networks (RNNs) can mimic any finite-state automaton in theory, and some workers demonstrated that this can hold in practice. We test the capability of generalized linear models, RCs, and Long Short-Term Memory (LSTM) RNN architectures to predict the stochastic processes generated by a large suite of probabilistic deterministic finite-state automata (PDFA). PDFAs provide an excellent performance benchmark in that they can be systematically enumerated, the randomness and correlation structure of their generated processes are exactly known, and their optimal memory-limited predictors are easily computed. Unsurprisingly, LSTMs outperform RCs, which outperform generalized linear models. Surprisingly, each of these methods can fall short of the maximal predictive accuracy by as much as 50% after training and, when optimized, tend to fall short of the maximal predictive accuracy by ~5%, even though previously available methods achieve maximal predictive accuracy with orders-of-magnitude less data. Thus, despite the representational universality of RCs and RNNs, using them can engender a surprising predictive gap for simple stimuli. One concludes that there is an important and underappreciated role for methods that infer "causal states" or "predictive state representations".