Acoustic Scene Classification (ASC) is a challenging task, as a single scene may involve multiple events that contain complex sound patterns. For example, a cooking scene may contain several sound sources including silverware clinking, chopping, frying, etc. What complicates ASC more is that classes of different activities could have overlapping sounds patterns (e.g. both cooking and dishwashing could have silverware clinking sound). In this paper, we propose a multi-head attention network to model the complex temporal input structures for ASC. The proposed network takes the audio's time-frequency representation as input, and it leverages standard VGG plus LSTM layers to extract high-level feature representation. Further more, it applies multiple attention heads to summarize various patterns of sound events into fixed dimensional representation, for the purpose of final scene classification. The whole network is trained in an end-to-end fashion with back-propagation. Experimental results confirm that our model discovers meaningful sound patterns through the attention mechanism, without using explicit supervision in the alignment. We evaluated our proposed model using DCASE 2018 Task 5 dataset, and achieved competitive performance on par with previous winner's results.

Transfer learning aims to learn robust classifiers for the target domain by leveraging knowledge from a source domain. Since the source and the target domains are usually from different distributions, existing methods mainly focus on adapting the cross-domain marginal or conditional distributions. However, in real applications, the marginal and conditional distributions usually have different contributions to the domain discrepancy. Existing methods fail to quantitatively evaluate the different importance of these two distributions, which will result in unsatisfactory transfer performance. In this paper, we propose a novel concept called Dynamic Distribution Adaptation (DDA), which is capable of quantitatively evaluating the relative importance of each distribution. DDA can be easily incorporated into the framework of structural risk minimization to solve transfer learning problems. On the basis of DDA, we propose two novel learning algorithms: (1) Manifold Dynamic Distribution Adaptation (MDDA) for traditional transfer learning, and (2) Dynamic Distribution Adaptation Network (DDAN) for deep transfer learning. Extensive experiments demonstrate that MDDA and DDAN significantly improve the transfer learning performance and setup a strong baseline over the latest deep and adversarial methods on digits recognition, sentiment analysis, and image classification. More importantly, it is shown that marginal and conditional distributions have different contributions to the domain divergence, and our DDA is able to provide good quantitative evaluation of their relative importance which leads to better performance. We believe this observation can be helpful for future research in transfer learning.

The electronic Schr\"odinger equation describes fundamental properties of molecules and materials, but cannot be solved exactly for larger systems than the hydrogen atom. Quantum Monte Carlo is a suitable method when high-quality approximations are sought, and its accuracy is in principle limited only by the flexibility of the used wave-function ansatz. Here we develop a deep-learning wave-function ansatz, dubbed PauliNet, which has the Hartree-Fock solution built in as a baseline, incorporates the physics of valid wave functions, and is trained using variational quantum Monte Carlo (VMC). Our deep-learning method achieves higher accuracy than comparable state-of-the-art VMC ansatzes for atoms, diatomic molecules and a strongly-correlated hydrogen chain. We anticipate that this method can reveal new physical insights and provide guidance for the design of molecules and materials where highly accurate quantum-mechanical solutions are needed, such as in transition metals and other strongly correlated systems.

Configuration spaces for computer systems can be challenging for traditional and automatic tuning strategies. Injecting task-specific knowledge into the tuner for a task may allow for more efficient exploration of candidate configurations. We apply this idea to the task of index set selection to accelerate database workloads. Index set selection has been amenable to recent applications of vanilla deep RL, but real deployments remain out of reach. In this paper, we explore how learning index selection can be enhanced with task-specific inductive biases, specifically by encoding these inductive biases in better action structures. Index selection-specific action representations arise when the problem is reformulated in terms of permutation learning and we rely on recent work for learning RL policies on permutations. Through this approach, we build an indexing agent that is able to achieve improved indexing and validate its behavior with task-specific statistics. Early experiments reveal that our agent can find configurations that are up to 40% smaller for the same levels of latency as compared with other approaches and indicate more intuitive indexing behavior.

As a result of social network popularity, in recent years, hate speech phenomenon has significantly increased. Due to its harmful effect on minority groups as well as on large communities, there is a pressing need for hate speech detection and filtering. However, automatic approaches shall not jeopardize free speech, so they shall accompany their decisions with explanations and assessment of uncertainty. Thus, there is a need for predictive machine learning models that not only detect hate speech but also help users understand when texts cross the line and become unacceptable. The reliability of predictions is usually not addressed in text classification. We fill this gap by proposing the adaptation of deep neural networks that can efficiently estimate prediction uncertainty. To reliably detect hate speech, we use Monte Carlo dropout regularization, which mimics Bayesian inference within neural networks. We evaluate our approach using different text embedding methods. We visualize the reliability of results with a novel technique that aids in understanding the classification reliability and errors.

Graph Convolutional Networks (GCN) have given the ability to model complex spatial and temporal dependencies in traffic data and improve the performance of predictions. In many studies, however, features that can represent the transportation networks such as speed limit, distance, and flow direction are overlooked. Learning without these structural features may not capture spatial dependencies and lead to low performance especially on roads with unusual characteristics. To address this challenge, we suggest a novel GCN structure that can incorporate multiple weights at the same time. The proposed model, Multi-Weight Traffic Graph Convolutional Networks (MW-TGC) conduct convolution operation on traffic data with multiple weighted adjacency matrices and combines the features obtained from each operation. The spatially isolated dimension reduction operation is conducted on the combined features to learn the dependencies among the features and reduce the size of output to a computationally feasible level. The output of multi-weight graph convolution is given to the Long Short-Term Memory (LSTM) to learn temporal dependencies. Experiment on two real-world datasets for 5min average speed of Seoul is conducted to evaluate the performance. The result shows that the proposed model outperforms the state-of-the-art models and reduces the inconsistency of prediction among roads with different characteristics.

Many problems in real life can be converted to combinatorial optimization problems (COPs) on graphs, that is to find a best node state configuration or a network structure such that the designed objective function is optimized under some constraints. However, these problems are notorious for their hardness to solve because most of them are NP-hard or NP-complete. Although traditional general methods such as simulated annealing (SA), genetic algorithms (GA) and so forth have been devised to these hard problems, their accuracy and time consumption are not satisfying in practice. In this work, we proposed a simple, fast, and general algorithm framework called Gumbel-softmax Optimization (GSO) for COPs. By introducing Gumbel-softmax technique which is developed in machine learning community, we can optimize the objective function directly by gradient descent algorithm regardless of the discrete nature of variables. We test our algorithm on four different problems including Sherrington-Kirkpatrick (SK) model, maximum independent set (MIS) problem, modularity optimization, and structural optimization problem. High-quality solutions can be obtained with much less time consuming compared to traditional approaches.

We propose a novel Shapley value approach to help address neural networks' interpretability and "vanishing gradient" problems. Our method is based on an accurate analytical approximation to the Shapley value of a neuron with ReLU activation. This analytical approximation admits a linear propagation of relevance across neural network layers, resulting in a simple, fast and sensible interpretation of neural networks' decision making process. We then derived a globally continuous and non-vanishing Shapley gradient, which can replace the conventional gradient in training neural network layers with ReLU activation, and leading to better training performance. We further derived a Shapley Activation (SA) function, which is a close approximation to ReLU but features the Shapley gradient. The SA is easy to implement in existing machine learning frameworks. Numerical tests show that SA consistently outperforms ReLU in training convergence, accuracy and stability.

--For many analytical problems the challenge is to handle huge amounts of available data. However, there are data science application areas where collecting information is difficult and costly, e.g., in the study of geological phenomena, rare diseases, faults in complex systems, insurance frauds, etc. In many such cases, generators of synthetic data with the same statistical and predictive properties as the actual data allow efficient simulations and development of tools and applications. In this work, we propose the incorporation of Monte Carlo Dropout method within Autoencoder (MCD-AE) and V ariational Autoencoder (MCD-V AE) as efficient generators of synthetic data sets. As the V ariational Autoencoder (V AE) is one of the most popular generator techniques, we explore its similarities and differences to the proposed methods. We compare the generated data sets with the original data based on statistical properties, structural similarity, and predictive similarity. The results obtained show a strong similarity between the results of V AE, MCD-V AE and MCD-AE; however, the proposed methods are faster and can generate values similar to specific selected initial instances. We live in times of big data; yet, there are many application areas that lack sufficient data for analyses, simulations, and development of analytical approaches.

Image captioning models are usually evaluated on their ability to describe a held-out set of images, not on their ability to generalize to unseen concepts. We study the problem of compositional generalization, which measures how well a model composes unseen combinations of concepts when describing images. State-of-the-art image captioning models show poor generalization performance on this task. We propose a multi-task model to address the poor performance, that combines caption generation and image--sentence ranking, and uses a decoding mechanism that re-ranks the captions according their similarity to the image. This model is substantially better at generalizing to unseen combinations of concepts compared to state-of-the-art captioning models.