Window-based Sentence Boundary Evaluation (WiSeBE)
Sentence Boundary Detection (SBD) has been a major research topic since Automatic Speech Recognition transcripts have been used for further Natural Language Processing tasks like Part of Speech Tagging, Question Answering or Automatic Summarization. But what about evaluation? Do standard evaluation metrics like precision, recall, F-score or classification error; and more important, evaluating an automatic system against a unique reference is enough to conclude how well a SBD system is performing given the final application of the transcript? In this paper we propose Window-based Sentence Boundary Evaluation (WiSeBE), a semi-supervised metric for evaluating Sentence Boundary Detection systems based on multi-reference (dis)agreement. We evaluate and compare the performance of different SBD systems over a set of Youtube transcripts using WiSeBE and standard metrics. This double evaluation gives an understanding of how WiSeBE is a more reliable metric for the SBD task. …

Deep Clustering
In the context of recent deep clustering studies, discriminative models dominate the literature and report the most competitive performances. These models learn a deep discriminative neural network classifier in which the labels are latent. Typically, they use multinomial logistic regression posteriors and parameter regularization, as is very common in supervised learning. It is generally acknowledged that discriminative objective functions (e.g., those based on the mutual information or the KL divergence) are more flexible than generative approaches (e.g., K-means) in the sense that they make fewer assumptions about the data distributions and, typically, yield much better unsupervised deep learning results. On the surface, several recent discriminative models may seem unrelated to K-means. This study shows that these models are, in fact, equivalent to K-means under mild conditions and common posterior models and parameter regularization. We prove that, for the commonly used logistic regression posteriors, maximizing the $L_2$ regularized mutual information via an approximate alternating direction method (ADM) is equivalent to a soft and regularized K-means loss. Our theoretical analysis not only connects directly several recent state-of-the-art discriminative models to K-means, but also leads to a new soft and regularized deep K-means algorithm, which yields competitive performance on several image clustering benchmarks. …

MLog
We demonstrate MLOG, a high-level language that integrates machine learning into data management systems. Unlike existing machine learning frameworks (e.g., TensorFlow, Theano, and Caffe), MLOG is declarative, in the sense that the system manages all data movement, data persistency, and machine-learning related optimizations (such as data batching) automatically. Our interactive demonstration will show audience how this is achieved based on the novel notion of tensoral views (TViews), which are similar to relational views but operate over tensors with linear algebra. With MLOG, users can succinctly specify not only simple models such as SVM (in just two lines), but also sophisticated deep learning models that are not supported by existing in-database analytics systems (e.g., MADlib, PAL, and SciDB), as a series of cascaded TViews. Given the declarative nature of MLOG, we further demonstrate how query/program optimization techniques can be leveraged to translate MLOG programs into native TensorFlow programs. The performance of the automatically generated Tensor- Flow programs is comparable to that of hand-optimized ones. …

Hardness-Aware Deep Metric Learning
This paper presents a hardness-aware deep metric learning (HDML) framework. Most previous deep metric learning methods employ the hard negative mining strategy to alleviate the lack of informative samples for training. However, this mining strategy only utilizes a subset of training data, which may not be enough to characterize the global geometry of the embedding space comprehensively. To address this problem, we perform linear interpolation on embeddings to adaptively manipulate their hard levels and generate corresponding label-preserving synthetics for recycled training, so that information buried in all samples can be fully exploited and the metric is always challenged with proper difficulty. Our method achieves very competitive performance on the widely used CUB-200-2011, Cars196, and Stanford Online Products datasets. …

Star-Transformer
Although the fully-connected attention-based model Transformer has achieved great successes on many NLP tasks, it has heavy structure and usually requires large training data. In this paper, we present the Star-Transformer, an alternative and light-weighted model of the Transformer. To reduce the model complexity, we replace the fully-connected structure with a star-shaped structure, in which every two non-adjacent nodes are connected through a shared relay node. Thus, the Star-Transformer has lower complexity than the standard Transformer (from quadratic to linear according to the input length) and preserves the ability to handle with the long-range dependencies. The experiments on four tasks (22 datasets) show the Star-Transformer achieved significant improvements against the standard Transformer for the modestly sized datasets. …

MATCHA
The trade-off between convergence error and communication delays in decentralized stochastic gradient descent~(SGD) is dictated by the sparsity of the inter-worker communication graph. In this paper, we propose MATCHA, a decentralized SGD method where we use matching decomposition sampling of the base graph to parallelize inter-worker information exchange so as to significantly reduce communication delay. At the same time, under standard assumptions for any general topology, in spite of the significant reduction of the communication delay, MATCHA maintains the same convergence rate as that of the state-of-the-art in terms of epochs. Experiments on a suite of datasets and deep neural networks validate the theoretical analysis and demonstrate the effectiveness of the proposed scheme as far as reducing communication delays is concerned. …

Gated Fully Fusion (GFF)
Semantic segmentation generates comprehensive understanding of scenes at a semantic level through densely predicting the category for each pixel. High-level features from Deep Convolutional Neural Networks already demonstrate their effectiveness in semantic segmentation tasks, however the coarse resolution of high-level features often leads to inferior results for small/thin objects where detailed information is important but missing. It is natural to consider importing low level features to compensate the lost detailed information in high level representations. Unfortunately, simply combining multi-level features is less effective due to the semantic gap existing among them. In this paper, we propose a new architecture, named Gated Fully Fusion(GFF), to selectively fuse features from multiple levels using gates in a fully connected way. Specifically, features at each level are enhanced by higher-level features with stronger semantics and lower-level features with more details, and gates are used to control the propagation of useful information which significantly reduces the noises during fusion. We achieve the state of the art results on two challenging scene understanding datasets, i.e., 82.3% mIoU on Cityscapes test set and 45.3% mIoU on ADE20K validation set. Codes and the trained models will be made publicly available. …

Centralized Kalman-Filtering (CKF)
We consider the Kalman-filtering problem with multiple sensors which are connected through a communication network. If all measurements are delivered to one place called fusion center and processed together, we call the process centralized Kalman-filtering (CKF). When there is no fusion center, each sensor can also solve the problem by using local measurements and exchanging information with its neighboring sensors, which is called distributed Kalman-filtering (DKF). Noting that CKF problem is a maximum likelihood estimation problem, which is a quadratic optimization problem, we reformulate DKF problem as a consensus optimization problem, resulting in that DKF problem can be solved by many existing distributed optimization algorithms. A new DKF algorithm employing the distributed dual ascent method is provided and its performance is evaluated through numerical experiments. …

KITTI Benchmark
We take advantage of our autonomous driving platform Annieway to develop novel challenging real-world computer vision benchmarks. Our tasks of interest are: stereo, optical flow, visual odometry, 3D object detection and 3D tracking. For this purpose, we equipped a standard station wagon with two high-resolution color and grayscale video cameras. Accurate ground truth is provided by a Velodyne laser scanner and a GPS localization system. Our datsets are captured by driving around the mid-size city of Karlsruhe, in rural areas and on highways. Up to 15 cars and 30 pedestrians are visible per image. Besides providing all data in raw format, we extract benchmarks for each task. For each of our benchmarks, we also provide an evaluation metric and this evaluation website. Preliminary experiments show that methods ranking high on established benchmarks such as Middlebury perform below average when being moved outside the laboratory to the real world. Our goal is to reduce this bias and complement existing benchmarks by providing real-world benchmarks with novel difficulties to the community. …

AlphaSeq
Sequences play an important role in many applications and systems. Discovering sequences with desired properties has long been an interesting intellectual pursuit. This paper puts forth a new paradigm, AlphaSeq, to discover desired sequences algorithmically using deep reinforcement learning (DRL) techniques. AlphaSeq treats the sequence discovery problem as an episodic symbol-filling game, in which a player fills symbols in the vacant positions of a sequence set sequentially during an episode of the game. Each episode ends with a completely-filled sequence set, upon which a reward is given based on the desirability of the sequence set. AlphaSeq models the game as a Markov Decision Process (MDP), and adapts the DRL framework of AlphaGo to solve the MDP. Sequences discovered improve progressively as AlphaSeq, starting as a novice, learns to become an expert game player through many episodes of game playing. Compared with traditional sequence construction by mathematical tools, AlphaSeq is particularly suitable for problems with complex objectives intractable to mathematical analysis. We demonstrate the searching capabilities of AlphaSeq in two applications: 1) AlphaSeq successfully rediscovers a set of ideal complementary codes that can zero-force all potential interferences in multi-carrier CDMA systems. 2) AlphaSeq discovers new sequences that triple the signal-to-interference ratio — benchmarked against the well-known Legendre sequence — of a mismatched filter estimator in pulse compression radar systems. …

Margin Disparity Discrepancy
This paper addresses the problem of unsupervised domain adaption from theoretical and algorithmic perspectives. Existing domain adaptation theories naturally imply minimax optimization algorithms, which connect well with the adversarial-learning based domain adaptation methods. However, several disconnections still form the gap between theory and algorithm. We extend previous theories (Ben-David et al., 2010; Mansour et al., 2009c) to multiclass classification in domain adaptation, where classifiers based on scoring functions and margin loss are standard algorithmic choices. We introduce a novel measurement, margin disparity discrepancy, that is tailored both to distribution comparison with asymmetric margin loss, and to minimax optimization for easier training. Using this discrepancy, we derive new generalization bounds in terms of Rademacher complexity. Our theory can be seamlessly transformed into an adversarial learning algorithm for domain adaptation, successfully bridging the gap between theory and algorithm. A series of empirical studies show that our algorithm achieves the state-of-the-art accuracies on challenging domain adaptation tasks. …

Quantized Generative Adversarial Network (QGAN)
The intensive computation and memory requirements of generative adversarial neural networks (GANs) hinder its real-world deployment on edge devices such as smartphones. Despite the success in model reduction of CNNs, neural network quantization methods have not yet been studied on GANs, which are mainly faced with the issues of both the effectiveness of quantization algorithms and the instability of training GAN models. In this paper, we start with an extensive study on applying existing successful methods to quantize GANs. Our observation reveals that none of them generates samples with reasonable quality because of the underrepresentation of quantized values in model weights, and the generator and discriminator networks show different sensitivities upon quantization methods. Motivated by these observations, we develop a novel quantization method for GANs based on EM algorithms, named as QGAN. We also propose a multi-precision algorithm to help find the optimal number of bits of quantized GAN models in conjunction with corresponding result qualities. Experiments on CIFAR-10 and CelebA show that QGAN can quantize GANs to even 1-bit or 2-bit representations with results of quality comparable to original models. …

Bayesian Hypergraph
We propose a directed acyclic hypergraph framework for a probabilistic graphical model that we call Bayesian hypergraphs. The space of directed acyclic hypergraphs is much larger than the space of chain graphs. Hence Bayesian hypergraphs can model much finer factorizations than Bayesian networks or LWF chain graphs and provide simpler and more computationally efficient procedures for factorizations and interventions. Bayesian hypergraphs also allow a modeler to represent causal patterns of interaction such as Noisy-OR graphically (without additional annotations). We introduce global, local and pairwise Markov properties of Bayesian hypergraphs and prove under which conditions they are equivalent. We define a projection operator, called shadow, that maps Bayesian hypergraphs to chain graphs, and show that the Markov properties of a Bayesian hypergraph are equivalent to those of its corresponding chain graph. We extend the causal interpretation of LWF chain graphs to Bayesian hypergraphs and provide corresponding formulas and a graphical criterion for intervention. …

MineRL Competition
Though deep reinforcement learning has led to breakthroughs in many difficult domains, these successes have required an ever-increasing number of samples. As state-of-the-art reinforcement learning (RL) systems require an exponentially increasing number of samples, their development is restricted to a continually shrinking segment of the AI community. Likewise, many of these systems cannot be applied to real-world problems, where environment samples are expensive. Resolution of these limitations requires new, sample-efficient methods. To facilitate research in this direction, we introduce the MineRL Competition on Sample Efficient Reinforcement Learning using Human Priors. The primary goal of the competition is to foster the development of algorithms which can efficiently leverage human demonstrations to drastically reduce the number of samples needed to solve complex, hierarchical, and sparse environments. To that end, we introduce: (1) the Minecraft ObtainDiamond task, a sequential decision making environment requiring long-term planning, hierarchical control, and efficient exploration methods; and (2) the MineRL-v0 dataset, a large-scale collection of over 60 million state-action pairs of human demonstrations that can be resimulated into embodied trajectories with arbitrary modifications to game state and visuals. Participants will compete to develop systems which solve the ObtainDiamond task with a limited number of samples from the environment simulator, Malmo. The competition is structured into two rounds in which competitors are provided several paired versions of the dataset and environment with different game textures. At the end of each round, competitors will submit containerized versions of their learning algorithms and they will then be trained/evaluated from scratch on a hold-out dataset-environment pair for a total of 4-days on a prespecified hardware platform. …

Compact Description
In critical applications of anomaly detection including computer security and fraud prevention, the anomaly detector must be configurable by the analyst to minimize the effort on false positives. One important way to configure the anomaly detector is by providing true labels for a few instances. We study the problem of label-efficient active learning to automatically tune anomaly detection ensembles and make four main contributions. First, we present an important insight into how anomaly detector ensembles are naturally suited for active learning. This insight allows us to relate the greedy querying strategy to uncertainty sampling, with implications for label-efficiency. Second, we present a novel formalism called compact description to describe the discovered anomalies and show that it can also be employed to improve the diversity of the instances presented to the analyst without loss in the anomaly discovery rate. Third, we present a novel data drift detection algorithm that not only detects the drift robustly, but also allows us to take corrective actions to adapt the detector in a principled manner. Fourth, we present extensive experiments to evaluate our insights and algorithms in both batch and streaming settings. Our results show that in addition to discovering significantly more anomalies than state-of-the-art unsupervised baselines, our active learning algorithms under the streaming-data setup are competitive with the batch setup. …

Header, Dictionary, Triples (HDT)
Currently RDF data is stored and sent in very verbose textual serialization formats that waste a lot of bandwidth and are expensive to parse and index. If RDF is meant to be machine understandable, why not use an appropriate format for that? HDT (Header, Dictionary, Triples) is a compact data structure and binary serialization format for RDF that keeps big datasets compressed to save space while maintaining search and browse operations without prior decompression. This makes it an ideal format for storing and sharing RDF datasets on the Web. …

Coded Partial Gradient Computation (CPGC)
Coded computation techniques provide robustness against straggling servers in distributed computing, with the following limitations: First, they increase decoding complexity. Second, they ignore computations carried out by straggling servers; and they are typically designed to recover the full gradient, and thus, cannot provide a balance between the accuracy of the gradient and per-iteration completion time. Here we introduce a hybrid approach, called coded partial gradient computation (CPGC), that benefits from the advantages of both coded and uncoded computation schemes, and reduces both the computation time and decoding complexity. …

Asynchronous Distributed Gibbs (ADG)
Gibbs sampling is a widely used Markov Chain Monte Carlo (MCMC) method for numerically approximating integrals of interest in Bayesian statistics and other mathematical sciences. It is widely believed that MCMC methods do not extend easily to parallel implementations, as their inherently sequential nature incurs a large synchronization cost. This means that new solutions are needed to bring Bayesian analysis fully into the era of large-scale computation. In this paper, we present a novel scheme – Asynchronous Distributed Gibbs (ADG) sampling – that allows us to perform MCMC in a parallel fashion with no synchronization or locking, avoiding the typical performance bottlenecks of parallel algorithms. Our method is especially attractive in settings, such as hierarchical random-effects modeling in which each observation has its own random effect, where the problem dimension grows with the sample size. We prove convergence under some basic regularity conditions, and discuss the proof for similar parallelization schemes for other iterative algorithms. We provide three examples that illustrate some of the algorithm’s properties with respect to scaling. Because our hardware resources are bounded, we have not yet found a limit to the algorithm’s scaling, and thus its true capabilities remain unknown. …

Evaluating Quantitative Understanding Aptitude in Textual Entailment (EQUATE)
Quantitative reasoning is an important component of reasoning that any intelligent natural language understanding system can reasonably be expected to handle. We present EQUATE (Evaluating Quantitative Understanding Aptitude in Textual Entailment), a new dataset to evaluate the ability of models to reason with quantities in textual entailment (including not only arithmetic and algebraic computation, but also other phenomena such as range comparisons and verbal reasoning with quantities). The average performance of 7 published textual entailment models on EQUATE does not exceed a majority class baseline, indicating that current models do not implicitly learn to reason with quantities. We propose a new baseline Q-REAS that manipulates quantities symbolically, achieving some success on numerical reasoning, but struggling at more verbal aspects of the task. We hope our evaluation framework will support the development of new models of quantitative reasoning in language understanding. …

Deep Weibull Model (DW-RNN)
One of the key challenges in predictive maintenance is to predict the impending downtime of an equipment with a reasonable prediction horizon so that countermeasures can be put in place. Classically, this problem has been posed in two different ways which are typically solved independently: (1) Remaining useful life (RUL) estimation as a long-term prediction task to estimate how much time is left in the useful life of the equipment and (2) Failure prediction (FP) as a short-term prediction task to assess the probability of a failure within a pre-specified time window. As these two tasks are related, performing them separately is sub-optimal and might results in inconsistent predictions for the same equipment. In order to alleviate these issues, we propose two methods: Deep Weibull model (DW-RNN) and multi-task learning (MTL-RNN). DW-RNN is able to learn the underlying failure dynamics by fitting Weibull distribution parameters using a deep neural network, learned with a survival likelihood, without training directly on each task. While DW-RNN makes an explicit assumption on the data distribution, MTL-RNN exploits the implicit relationship between the long-term RUL and short-term FP tasks to learn the underlying distribution. Additionally, both our methods can leverage the non-failed equipment data for RUL estimation. We demonstrate that our methods consistently outperform baseline RUL methods that can be used for FP while producing consistent results for RUL and FP. We also show that our methods perform at par with baselines trained on the objectives optimized for either of the two tasks. …

Bayesian Conditional Generative Adverserial Networks (BC-GAN)
Traditional GANs use a deterministic generator function (typically a neural network) to transform a random noise input $z$ to a sample $\mathbf{x}$ that the discriminator seeks to distinguish. We propose a new GAN called Bayesian Conditional Generative Adversarial Networks (BC-GANs) that use a random generator function to transform a deterministic input $y’$ to a sample $\mathbf{x}$. Our BC-GANs extend traditional GANs to a Bayesian framework, and naturally handle unsupervised learning, supervised learning, and semi-supervised learning problems. Experiments show that the proposed BC-GANs outperforms the state-of-the-arts. …

Generalized Canonical Polyadic Tensor Decomposition (GCP)
Tensor decomposition is a fundamental unsupervised machine learning method in data science, with applications including network analysis and sensor data processing. This work develops a generalized canonical polyadic (GCP) low-rank tensor decomposition that allows other loss functions besides squared error. For instance, we can use logistic loss or Kullback-Leibler divergence, enabling tensor decomposition for binary or count data. We present a variety statistically-motivated loss functions for various scenarios. We provide a generalized framework for computing gradients and handling missing data that enables the use of standard optimization methods for fitting the model. We demonstrate the flexibility of GCP on several real-world examples including interactions in a social network, neural activity in a mouse, and monthly rainfall measurements in India. …

Mutex Watershed
Image partitioning, or segmentation without semantics, is the task of decomposing an image into distinct segments, or equivalently to detect closed contours. Most prior work either requires seeds, one per segment; or a threshold; or formulates the task as multicut / correlation clustering, an NP-hard problem. Here, we propose a greedy algorithm for signed graph partitioning, the ‘Mutex Watershed’. Unlike seeded watershed, the algorithm can accommodate not only attractive but also repulsive cues, allowing it to find a previously unspecified number of segments without the need for explicit seeds or a tunable threshold. We also prove that this simple algorithm solves to global optimality an objective function that is intimately related to the multicut / correlation clustering integer linear programming formulation. The algorithm is deterministic, very simple to implement, and has empirically linearithmic complexity. When presented with short-range attractive and long-range repulsive cues from a deep neural network, the Mutex Watershed gives the best results currently known for the competitive ISBI 2012 EM segmentation benchmark. …

Action-Specific Deep Recurrent Q-Network (ADRQN)
Deep Reinforcement Learning (RL) recently emerged as one of the most competitive approaches for learning in sequential decision making problems with fully observable environments, e.g., computer Go. However, very little work has been done in deep RL to handle partially observable environments. We propose a new architecture called Action-specific Deep Recurrent Q-Network (ADRQN) to enhance learning performance in partially observable domains. Actions are encoded by a fully connected layer and coupled with a convolutional observation to form an action-observation pair. The time series of action-observation pairs are then integrated by an LSTM layer that learns latent states based on which a fully connected layer computes Q-values as in conventional Deep Q-Networks (DQNs). We demonstrate the effectiveness of our new architecture in several partially observable domains, including flickering Atari games. …

Boosting Smooth Transition Regression Tree (BooST)
In this paper we introduce a new machine learning (ML) model for nonlinear regression called Boosting Smooth Transition Regression Tree (BooST). The main advantage of the BooST is that it estimates the derivatives (partial effects) of very general nonlinear models, providing more interpretation than other tree based models concerning the mapping between the covariates and the dependent variable. We provide some asymptotic theory that shows consistency of the partial derivatives and we present some examples on simulated and empirical data. …

Segmented Linear Regression
Segmented linear regression with two segments separated by a breakpoint can be useful to quantify an abrupt change of the response function (Yr) of a varying influential factor (x). The breakpoint can be interpreted as a critical, safe, or threshold value beyond or below which (un)desired effects occur. The breakpoint can be important in decision making. …

MoCHIN
As one type of complex networks widely-seen in real-world application, heterogeneous information networks (HINs) often encapsulate higher-order interactions that crucially reflect the complex nature among nodes and edges in real-world data. Modeling higher-order interactions in HIN facilitates the user-guided clustering problem by providing an informative collection of signals. At the same time, network motifs have been used extensively to reveal higher-order interactions and network semantics in homogeneous networks. Thus, it is natural to extend the use of motifs to HIN, and we tackle the problem of user-guided clustering in HIN by using motifs. We highlight the benefits of comprehensively modeling higher-order interactions instead of decomposing the complex relationships to pairwise interaction. We propose the MoCHIN model which is applicable to arbitrary forms of HIN motifs, which is often necessary for the application scenario in HINs due to their rich and diverse semantics encapsulated in the heterogeneity. To overcome the curse of dimensionality since the tensor size grows exponentially as the number of nodes increases in our model, we propose an efficient inference algorithm for MoCHIN. In our experiment, MoCHIN surpasses all baselines in three evaluation tasks under different metrics. The advantage of our model when the supervision is weak is also discussed in additional experiments. …

GaborNet
The article describes a system for image recognition using deep convolutional neural networks. Modified network architecture is proposed that focuses on improving convergence and reducing training complexity. The filters in the first layer of the network are constrained to fit the Gabor function. The parameters of Gabor functions are learnable and are updated by standard backpropagation techniques. The system was implemented on Python, tested on several datasets and outperformed the common convolutional networks. …

BNN+
Deep neural networks (DNN) are widely used in many applications. However, their deployment on edge devices has been difficult because they are resource hungry. Binary neural networks (BNN) help to alleviate the prohibitive resource requirements of DNN, where both activations and weights are limited to $1$-bit. We propose an improved binary training method (BNN+), by introducing a regularization function that encourages training weights around binary values. In addition to this, to enhance model performance we add trainable scaling factors to our regularization functions. Furthermore, we use an improved approximation of the derivative of the sign activation function in the backward computation. These additions are based on linear operations that are easily implementable into the binary training framework. We show experimental results on CIFAR-10 obtaining an accuracy of $86.7\%$, on AlexNet and $91.3\%$ with VGG network. On ImageNet, our method also outperforms the traditional BNN method and XNOR-net, using AlexNet by a margin of $4\%$ and $2\%$ top-$1$ accuracy respectively. …

Propensity Score
A propensity score is the probability of a unit (e.g., person, classroom, school) being assigned to a particular treatment given a set of observed covariates. Propensity scores are used to reduce selection bias by equating groups based on these covariates. …

Continual Learning via Neural Pruning (CLNP)
We introduce Continual Learning via Neural Pruning (CLNP), a new method aimed at lifelong learning in fixed capacity models based on neuronal model sparsification. In this method, subsequent tasks are trained using the inactive neurons and filters of the sparsified network and cause zero deterioration to the performance of previous tasks. In order to deal with the possible compromise between model sparsity and performance, we formalize and incorporate the concept of graceful forgetting: the idea that it is preferable to suffer a small amount of forgetting in a controlled manner if it helps regain network capacity and prevents uncontrolled loss of performance during the training of future tasks. CLNP also provides simple continual learning diagnostic tools in terms of the number of free neurons left for the training of future tasks as well as the number of neurons that are being reused. In particular, we see in experiments that CLNP verifies and automatically takes advantage of the fact that the features of earlier layers are more transferable. We show empirically that CLNP leads to significantly improved results over current weight elasticity based methods. …

Surrogate Variable Analysis (SVA)
Modern high-throughput molecular biology experiments measure data for thousands of related features and seek to rank those features for association with some variables of experimental or clinical importance. The process of ranking features for association with primary variables is complicated by genetic, environmental, and technical factors that influence hundreds or thousands of features at a time. In highdimensional experiments these factors are often unknown, unmeasured, or incapable of being tractably modeled. Consistent patterns of variation across features due to unmeasured or unmodeled factors can confound the relationship between the primary variables and the measured features. In this thesis we provide a statistical framework for modeling large-scale noise dependence caused by unmeasured or unmodeled factors in high-throughput data. We argue that estimating the sources of noise dependence is more appropriate than estimating the pairwise covariance between all features when the number of features is large. A direct connection is made with the well-studied problem of multiple testing dependence, which typically focuses on the distribution of P-values from multiple testing procedures. We introduce the concept of surrogate variables, estimable linear combinations of the true unmeasured or unmodeled factors causing noise dependence, that can be included when modeling the relationship between the primary variables and the feature level data. We also propose algorithms for estimating surrogate variables based on principal component analysis of relevant subsets of features. Under certain conditions accounting for the estimated surrogate variables asymptotically corrects the ranking and error rate estimation in high-throughput data analysis. We also discuss pathological situations when surrogate variables can not be estimated. To illustrate the power of this approach, we apply our estimates of the surrogate variables to improve reproducibility in a large clinical gene expression study of trauma related outcomes. …

Indirect Inference
Indirect inference is a simulation-based method for estimating the parameters of economic models. Its hallmark is the use of an auxiliary model to capture aspects of the data upon which to base the estimation. The parameters of the auxiliary model can be estimated using either the observed data or data simulated from the economic model. Indirect inference chooses the parameters of the economic model so that these two estimates of the parameters of the auxiliary model are as close as possible. The auxiliary model need not be correctly specified; when it is, indirect inference is equivalent to maximum likelihood. …

TSAVE
Supervised dimension reduction for time series is challenging as there may be temporal dependence between the response $y$ and the predictors $\boldsymbol x$. Recently a time series version of sliced inverse regression, TSIR, was suggested, which applies approximate joint diagonalization of several supervised lagged covariance matrices to consider the temporal nature of the data. In this paper we develop this concept further and propose a time series version of sliced average variance estimation, TSAVE. As both TSIR and TSAVE have their own advantages and disadvantages, we consider furthermore a hybrid version of TSIR and TSAVE. Based on examples and simulations we demonstrate and evaluate the differences between the three methods and show also that they are superior to apply their iid counterparts to when also using lagged values of the explaining variables as predictors. …

ImplicitCE
Although modern recommendation systems can exploit the structure in users’ item feedback, most are powerless in the face of new users who provide no structure for them to exploit. In this paper we introduce ImplicitCE, an algorithm for recommending items to new users during their sign-up flow. ImplicitCE works by transforming users’ implicit feedback towards auxiliary domain items into an embedding in the target domain item embedding space. ImplicitCE learns these embedding spaces and transformation function in an end-to-end fashion and can co-embed users and items with any differentiable similarity function. To train ImplicitCE we explore methods for maximizing the correlations between model predictions and users’ affinities and introduce Sample Correlation Update, a novel and extremely simple training strategy. Finally, we show that ImplicitCE trained with Sample Correlation Update outperforms a variety of state of the art algorithms and loss functions on both a large scale Twitter dataset and the DBLP dataset. …

CUR Decomposition
This article discusses a useful tool in dimensionality reduction and low-rank matrix approximation called the CUR decomposition. Various viewpoints of this method in the literature are synergized and are compared and contrasted, included in this is a new characterization of exact CUR decompositions. A novel perturbation analysis is performed on CUR approximations of noisy versions of low-rank matrices, which compares them with the putative CUR decomposition of the underlying low-rank part. Additionally, we give new column and row sampling results which allow one to conclude that a CUR decomposition of a low-rank matrix is attained with high probability. We then illustrate the stability of these sampling methods under the perturbations studied before, and provide numerical illustrations of the methods and bounds discussed. …

Concept-Oriented Deep Learning (CODL)
Concepts are the foundation of human deep learning, understanding, and knowledge integration and transfer. We propose concept-oriented deep learning (CODL) which extends (machine) deep learning with concept representations and conceptual understanding capability. CODL addresses some of the major limitations of deep learning: interpretability, transferability, contextual adaptation, and requirement for lots of labeled training data. We discuss the major aspects of CODL including concept graph, concept representations, concept exemplars, and concept representation learning systems supporting incremental and continual learning. …

EXTRA
State-of-the-art in network science of teams offers effective recommendation methods to answer questions like who is the best replacement, what is the best team expansion strategy, but lacks intuitive ways to explain why the optimization algorithm gives the specific recommendation for a given team optimization scenario. To tackle this problem, we develop an interactive prototype system, EXTRA, as the first step towards addressing such a sense-making challenge, through the lens of the underlying network where teams embed, to explain the team recommendation results. The main advantages are (1) Algorithm efficacy: we propose an effective and fast algorithm to explain random walk graph kernel, the central technique for networked team recommendation; (2) Intuitive visual explanation: we present intuitive visual analysis of the recommendation results, which can help users better understand the rationality of the underlying team recommendation algorithm. …

Rectified Decision Tree (ReDT)
How to obtain a model with good interpretability and performance has always been an important research topic. In this paper, we propose rectified decision trees (ReDT), a knowledge distillation based decision trees rectification with high interpretability, small model size, and empirical soundness. Specifically, we extend the impurity calculation and the pure ending condition of the classical decision tree to propose a decision tree extension that allows the use of soft labels generated by a well-trained teacher model in training and prediction process. It is worth noting that for the acquisition of soft labels, we propose a new multiple cross-validation based method to reduce the effects of randomness and overfitting. These approaches ensure that ReDT retains excellent interpretability and even achieves fewer nodes than the decision tree in the aspect of compression while having relatively good performance. Besides, in contrast to traditional knowledge distillation, back propagation of the student model is not necessarily required in ReDT, which is an attempt of a new knowledge distillation approach. Extensive experiments are conducted, which demonstrates the superiority of ReDT in interpretability, compression, and empirical soundness. …

HyperTrick
Training intelligent agents through reinforcement learning is a notoriously unstable procedure. Massive parallelization on GPUs and distributed systems has been exploited to generate a large amount of training experiences and consequently reduce instabilities, but the success of training remains strongly influenced by the choice of the hyperparameters. To overcome this issue, we introduce HyperTrick, a new metaoptimization algorithm, and show its effective application to tune hyperparameters in the case of deep reinforcement learning, while learning to play different Atari games on a distributed system. Our analysis provides evidence of the interaction between the identification of the optimal hyperparameters and the learned policy, that is typical of the case of metaoptimization for deep reinforcement learning. When compared with state-of-the-art metaoptimization algorithms, HyperTrick is characterized by a simpler implementation and it allows learning similar policies, while making a more effective use of the computational resources in a distributed system. …