The residual network is now one of the most effective structures in deep learning, which utilizes the skip connections to ``guarantee" the performance will not get worse. However, the non-convexity of the neural network makes it unclear whether the skip connections do provably improve the learning ability since the nonlinearity may create many local minima. In some previous works \cite{freeman2016topology}, it is shown that despite the non-convexity, the loss landscape of the two-layer ReLU network has good properties when the number $m$ of hidden nodes is very large. In this paper, we follow this line to study the topology (sub-level sets) of the loss landscape of deep ReLU neural networks with a skip connection and theoretically prove that the skip connection network inherits the good properties of the two-layer network and skip connections can help to control the connectedness of the sub-level sets, such that any local minima worse than the global minima of some two-layer ReLU network will be very ``shallow". The ``depth" of these local minima are at most $O(m^{(\eta-1)/n})$, where $n$ is the input dimension, $\eta<1$. This provides a theoretical explanation for the effectiveness of the skip connection in deep learning.

User behavior and feature interactions are crucial in deep learning-based recommender systems. There has been a diverse set of behavior modeling and interaction exploration methods in the literature. Nevertheless, the design of task-aware recommender systems still requires feature engineering and architecture engineering from domain experts. In this work, we introduce AMER, namely Automatic behavior Modeling and interaction Exploration in Recommender systems with Neural Architecture Search (NAS). The core contributions of AMER include the three-stage search space and the tailored three-step searching pipeline. In the first step, AMER searches for residual blocks that incorporate commonly used operations in the block-wise search space of stage 1 to model sequential patterns in user behavior. In the second step, it progressively investigates useful low-order and high-order feature interactions in the non-sequential interaction space of stage 2. Finally, an aggregation multi-layer perceptron (MLP) with shortcut connection is selected from flexible dimension settings of stage~3 to combine features extracted from the previous steps. For efficient and effective NAS, AMER employs the one-shot random search in all three steps. Further analysis reveals that AMER's search space could cover most of the representative behavior extraction and interaction investigation methods, which demonstrates the universality of our design. The extensive experimental results over various scenarios reveal that AMER could outperform competitive baselines with elaborate feature engineering and architecture engineering, indicating both effectiveness and robustness of the proposed method.

The success of deep supervised learning depends on its automatic data representation abilities. Among all the characteristics of an ideal representation for high-dimensional complex data, information preservation, low dimensionality and disentanglement are the most essential ones. In this work, we propose a deep dimension reduction (DDR) approach to achieving a good data representation with these characteristics for supervised learning. At the population level, we formulate the ideal representation learning task as finding a nonlinear dimension reduction map that minimizes the sum of losses characterizing conditional independence and disentanglement. We estimate the target map at the sample level nonparametrically with deep neural networks. We derive a bound on the excess risk of the deep nonparametric estimator. The proposed method is validated via comprehensive numerical experiments and real data analysis in the context of regression and classification.

Deep learning networks are typically trained by Stochastic Gradient Descent (SGD) methods that iteratively improve the model parameters by estimating a gradient on a very small fraction of the training data. A major roadblock faced when increasing the batch size to a substantial fraction of the training data for improving training time is the persistent degradation in performance (generalization gap). To address this issue, recent work propose to add small perturbations to the model parameters when computing the stochastic gradients and report improved generalization performance due to smoothing effects. However, this approach is poorly understood; it requires often model-specific noise and fine-tuning. To alleviate these drawbacks, we propose to use instead computationally efficient extrapolation (extragradient) to stabilize the optimization trajectory while still benefiting from smoothing to avoid sharp minima. This principled approach is well grounded from an optimization perspective and we show that a host of variations can be covered in a unified framework that we propose. We prove the convergence of this novel scheme and rigorously evaluate its empirical performance on ResNet, LSTM, and Transformer. We demonstrate that in a variety of experiments the scheme allows scaling to much larger batch sizes than before whilst reaching or surpassing SOTA accuracy.

Training and inference efficiency of deep neural networks highly rely on the performance of tensor operators on hardware platforms. Manually optimized tensor operators have limitations in terms of supporting new operators or supporting new hardware platforms. Therefore, automatically optimizing device code configurations of tensor operators is getting increasingly attractive. However, current methods for tensor operator optimization usually suffer from poor sample-efficiency due to the combinatorial search space. In this work, we propose a novel evolutionary method, OpEvo, which efficiently explores the search spaces of tensor operators by introducing a topology-aware mutation operation based on q-random walk distribution to leverage the topological structures over the search spaces. Our comprehensive experiment results show that OpEvo can find the best configuration with the least number of trials and the lowest variance compared with state-of-the-art methods. All code of this work is available online.

The LSTM network was proposed to overcome the difficulty in learning long-term dependence, and has made significant advancements in applications. With its success and drawbacks in mind, this paper raises the question - do RNN and LSTM have long memory? We answer it partially by proving that RNN and LSTM do not have long memory from a statistical perspective. A new definition for long memory networks is further introduced, and it requires the model weights to decay at a polynomial rate. To verify our theory, we convert RNN and LSTM into long memory networks by making a minimal modification, and their superiority is illustrated in modeling long-term dependence of various datasets.

A complex combination of simultaneous supervised-unsupervised learning is believed to be the key to humans performing tasks seamlessly across multiple domains or tasks. This phenomenon of cross-domain learning has been very well studied in domain adaptation literature. Recent domain adaptation works rely on an indirect way of first aligning the source and target domain distributions and then train a classifier on the labeled source domain to classify the target domain. However, this approach has the main drawback that obtaining a near-perfect alignment of the domains in itself might be difficult/impossible (e.g., language domains). To address this, we follow Vapnik's imperative of statistical learning that states any desired problem should be solved in the most direct way rather than solving a more general intermediate task and propose a direct approach to domain adaptation that does not require domain alignment. We propose a model referred Contradistinguisher that learns contrastive features and whose objective is to jointly learn to contradistinguish the unlabeled target domain in an unsupervised way and classify in a supervised way on the source domain. We achieve the state-of-the-art on Office-31 and VisDA-2017 datasets in both single-source and multi-source settings. We also demonstrate that the contradistinguish loss improves the model performance by increasing the shape bias.

Researchers in Italy have melded the emerging science of convolutional neural networks (CNNs) with deep learning -- a discipline within artificial intelligence -- to achieve a system of market forecasting with the potential for greater gains and fewer losses than previous attempts to use AI methods to manage stock portfolios. The team, led by Prof. Silvio Barra at the University of Cagliari, published their findings on IEEE/CAA Journal of Automatica Sinica. The University of Cagliari-based team set out to create an AI-managed "buy and hold" (B&H) strategy -- a system of deciding whether to take one of three possible actions -- a long action (buying a stock and selling it before the market closes), a short action (selling a stock, then buying it back before the market closes), and a hold (deciding not to invest in a stock that day). At the heart of their proposed system is an automated cycle of analyzing layered images generated from current and past market data. Older B&H systems based their decisions on machine learning, a discipline that leans heavily on predictions based on past performance.

With creative AI emerging, art creation doesn't seem to be unique to humans, not anymore. Creativity is one of the few traits that make humans different from other species. We alone can make music and art that speak to our experiences or illuminate truths about our world. But suddenly, humans' artistic abilities have some competition--and from a decidedly non-human source; Artificial Intelligence. Over the last couple of years, there have been some remarkable examples of art produced by deep learning algorithms.

The new approach is based on learning plausible face sketches from real face sketch images and finding the main aspects that approximate the input sketch. Treating input sketches as soft constraints allowed the researchers to produce high-quality face images with increased plausibility. The new approach is said to include three core modules - CE (Component Embedding), FM (Feature Mapping), and IS (Image Synthesis). The first one adopts an auto-encoder architecture and separately learns five feature descriptors ( left-eye, right-eye, nose, mouth, and remainder). The other two modules develop deep learning sub-network for conditional image generation.