PyTorchKR Papers

최신 논문

총 579편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2105.04983

Tensor-Train Recurrent Neural Networks for Interpretable Multi-Way Financial Forecasting

Recurrent Neural Networks (RNNs) represent the de facto standard machine learning tool for sequence modelling, owing to their expressive power and memory. However, when dealing with large dimensional data, the corresponding exponential increase in the number of parameters imposes a computational bottleneck. The necessity to equip RNNs with the ability to deal with the curse of dimensionality, such as through the parameter compression ability inherent to tensors, has led to the development of the Tensor-Train RNN (TT-RNN). Despite achieving promising results in many applications, the full potential of the TT-RNN is yet to be explored in the context of interpretable financial modelling, a notoriously challenging task characterized by multi-modal data with low signal-to-noise ratio. To address this issue, we investigate the potential of TT-RNN in the task of financial forecasting of currencies. We show, through the analysis of TT-factors, that the physical meaning underlying tensor decomposition, enables the TT-RNN model to aid the interpretability of results, thus mitigating the notorious "black-box" issue associated with neural networks. Furthermore, simulation results highlight the regularization power of TT decomposition, demonstrating the superior performance of TT-RNN over its uncompressed RNN counterpart and other tensor forecasting methods.

arXiv:2105.01751

Reconstruction Algorithms for Low-Rank Tensors and Depth-3 Multilinear Circuits

We give new and efficient black-box reconstruction algorithms for some classes of depth-$3$ arithmetic circuits. As a consequence, we obtain the first efficient algorithm for computing the tensor rank and for finding the optimal tensor decomposition as a sum of rank-one tensors when then input is a constant-rank tensor. More specifically, we provide efficient learning algorithms that run in randomized polynomial time over general fields and in deterministic polynomial time over the reals and the complex numbers for the following classes: (1) Set-multilinear depth-$3$ circuits of constant top fan-in $\Sigma\Pi\Sigma\{\sqcup_j X_j\}(k)$ circuits). As a consequence of our algorithm, we obtain the first polynomial time algorithm for tensor rank computation and optimal tensor decomposition of constant-rank tensors. This result holds for $d$ dimensional tensors for any $d$, but is interesting even for $d=3$. (2) Sums of powers of constantly many linear forms ($\Sigma\wedge\Sigma$ circuits). As a consequence we obtain the first polynomial-time algorithm for tensor rank computation and optimal tensor decomposition of constant-rank symmetric tensors. (3) Multilinear depth-3 circuits of constant top fan-in (multilinear $\Sigma\Pi\Sigma(k)$ circuits). Our algorithm works over all fields of characteristic 0 or large enough characteristic. Prior to our work the only efficient algorithms known were over polynomially-sized finite fields (see. Karnin-Shpilka 09'). Prior to our work, the only polynomial-time or even subexponential-time algorithms known (deterministic or randomized) for subclasses of $\Sigma\Pi\Sigma(k)$ circuits that also work over large/infinite fields were for the setting when the top fan-in $k$ is at most $2$ (see Sinha 16' and Sinha 20').

arXiv:2105.01136

Learning Good State and Action Representations via Tensor Decomposition

The transition kernel of a continuous-state-action Markov decision process (MDP) admits a natural tensor structure. This paper proposes a tensor-inspired unsupervised learning method to identify meaningful low-dimensional state and action representations from empirical trajectories. The method exploits the MDP's tensor structure by kernelization, importance sampling and low-Tucker-rank approximation. This method can be further used to cluster states and actions respectively and find the best discrete MDP abstraction. We provide sharp statistical error bounds for tensor concentration and the preservation of diffusion distance after embedding. We further prove that the learned state/action abstractions provide accurate approximations to latent block structures if they exist, enabling function approximation in downstream tasks such as policy evaluation.

arXiv:2105.01029공식 코드 · Papers with Code

Initialization and Regularization of Factorized Neural Layers

Factorized layers--operations parameterized by products of two or more matrices--occur in a variety of deep learning contexts, including compressed model training, certain types of knowledge distillation, and multi-head self-attention architectures. We study how to initialize and regularize deep nets containing such layers, examining two simple, understudied schemes, spectral initialization and Frobenius decay, for improving their performance. The guiding insight is to design optimization routines for these networks that are as close as possible to that of their well-tuned, non-decomposed counterparts; we back this intuition with an analysis of how the initialization and regularization schemes impact training with gradient descent, drawing on modern attempts to understand the interplay of weight-decay and batch-normalization. Empirically, we highlight the benefits of spectral initialization and Frobenius decay across a variety of settings. In model compression, we show that they enable low-rank methods to significantly outperform both unstructured sparsity and tensor methods on the task of training low-memory residual networks; analogs of the schemes also improve the performance of tensor decomposition techniques. For knowledge distillation, Frobenius decay enables a simple, overcomplete baseline that yields a compact model from over-parameterized training without requiring retraining with or pruning a teacher network. Finally, we show how both schemes applied to multi-head attention lead to improved performance on both translation and unsupervised pre-training.

arXiv:2104.14562

Stochastic Mirror Descent for Low-Rank Tensor Decomposition Under Non-Euclidean Losses

This work considers low-rank canonical polyadic decomposition (CPD) under a class of non-Euclidean loss functions that frequently arise in statistical machine learning and signal processing. These loss functions are often used for certain types of tensor data, e.g., count and binary tensors, where the least squares loss is considered unnatural.Compared to the least squares loss, the non-Euclidean losses are generally more challenging to handle. Non-Euclidean CPD has attracted considerable interests and a number of prior works exist. However, pressing computational and theoretical challenges, such as scalability and convergence issues, still remain. This work offers a unified stochastic algorithmic framework for large-scale CPD decomposition under a variety of non-Euclidean loss functions. Our key contribution lies in a tensor fiber sampling strategy-based flexible stochastic mirror descent framework. Leveraging the sampling scheme and the multilinear algebraic structure of low-rank tensors, the proposed lightweight algorithm ensures global convergence to a stationary point under reasonable conditions. Numerical results show that our framework attains promising non-Euclidean CPD performance. The proposed framework also exhibits substantial computational savings compared to state-of-the-art methods.

arXiv:2104.10625공식 코드 · Papers with Code

Searching to Sparsify Tensor Decomposition for N-ary Relational Data

Tensor, an extension of the vector and matrix to the multi-dimensional case, is a natural way to describe the N-ary relational data. Recently, tensor decomposition methods have been introduced into N-ary relational data and become state-of-the-art on embedding learning. However, the performance of existing tensor decomposition methods is not as good as desired. First, they suffer from the data-sparsity issue since they can only learn from the N-ary relational data with a specific arity, i.e., parts of common N-ary relational data. Besides, they are neither effective nor efficient enough to be trained due to the over-parameterization problem. In this paper, we propose a novel method, i.e., S2S, for effectively and efficiently learning from the N-ary relational data. Specifically, we propose a new tensor decomposition framework, which allows embedding sharing to learn from facts with mixed arity. Since the core tensors may still suffer from the over-parameterization, we propose to reduce parameters by sparsifying the core tensors while retaining their expressive power using neural architecture search (NAS) techniques, which can search for data-dependent architectures. As a result, the proposed S2S not only guarantees to be expressive but also efficiently learns from mixed arity. Finally, empirical results have demonstrated that S2S is efficient to train and achieves state-of-the-art performance.

arXiv:0711.2023공식 코드 · Papers with Code

Empirical Evaluation of Four Tensor Decomposition Algorithms

Higher-order tensor decompositions are analogous to the familiar Singular Value Decomposition (SVD), but they transcend the limitations of matrices (second-order tensors). SVD is a powerful tool that has achieved impressive results in information retrieval, collaborative filtering, computational linguistics, computational vision, and other fields. However, SVD is limited to two-dimensional arrays of data (two modes), and many potential applications have three or more modes, which require higher-order tensor decompositions. This paper evaluates four algorithms for higher-order tensor decomposition: Higher-Order Singular Value Decomposition (HO-SVD), Higher-Order Orthogonal Iteration (HOOI), Slice Projection (SP), and Multislice Projection (MP). We measure the time (elapsed run time), space (RAM and disk space requirements), and fit (tensor reconstruction accuracy) of the four algorithms, under a variety of conditions. We find that standard implementations of HO-SVD and HOOI do not scale up to larger tensors, due to increasing RAM requirements. We recommend HOOI for tensors that are small enough for the available RAM and MP for larger tensors.

arXiv:2104.08580공식 코드 · Papers with Code

Uncovering audio patterns in music with Nonnegative Tucker Decomposition for structural segmentation

Recent work has proposed the use of tensor decomposition to model repetitions and to separate tracks in loop-based electronic music. The present work investigates further on the ability of Nonnegative Tucker Decompositon (NTD) to uncover musical patterns and structure in pop songs in their audio form. Exploiting the fact that NTD tends to express the content of bars as linear combinations of a few patterns, we illustrate the ability of the decomposition to capture and single out repeated motifs in the corresponding compressed space, which can be interpreted from a musical viewpoint. The resulting features also turn out to be efficient for structural segmentation, leading to experimental results on the RWC Pop data set which are potentially challenging state-of-the-art approaches that rely on extensive example-based learning schemes.

arXiv:2007.01949

On the use of higher-order tensors to model muscle synergies

The muscle synergy concept provides the best framework to understand motor control and it has been recently utilised in many applications such as prosthesis control. The current muscle synergy model relies on decomposing multi-channel surface Electromyography (EMG) signals into a synergy matrix (spatial mode) and its weighting function (temporal mode). This is done using several matrix factorisation techniques, with Non-negative matrix factorisation (NMF) being the most prominent method. Here, we introduce a 4th-order tensor muscle synergy model that extends the current state of the art by taking spectral information and repetitions (movements) into account. This adds more depth to the model and provides more synergistic information. In particular, we illustrate a proof-of-concept study where the Tucker3 tensor decomposition model was applied to a subset of wrist movements from the Ninapro database. The results showed the potential of Tucker3 tensor factorisation in finding patterns of muscle synergies with information about the movements and highlights the differences between the current and proposed model.

arXiv:2007.01944

Consistency of Muscle Synergies Extracted via Higher-Order Tensor Decomposition Towards Myoelectric Control

In recent years, muscle synergies have been pro-posed for proportional myoelectric control. Synergies were extracted using matrix factorisation techniques (mainly non-negative matrix factorisation, NMF), which requires identification of synergies to tasks or movements. In addition, NMF methods were viable only with a task dimension of 2 degrees of freedoms(DoFs). Here, the potential use of a higher-order tensor model for myoelectric control is explored. We assess the ability of a constrained Tucker tensor decomposition to estimate consistent synergies when the task dimensionality is increased up to 3-DoFs. Synergies extracted from 3rd-order tensor of 1 and 3 DoFs were compared. Results showed that muscle synergies extracted via constrained Tucker decomposition were consistent with the increase of task-dimension. Hence, these results support the consideration of proportional 3-DoF myoelectric control based on tensor decompositions.

arXiv:2101.09672

Towards Overfitting Avoidance: Tuning-free Tensor-aided Multi-user Channel Estimation for 3D Massive MIMO Communications

Channel estimation has long been deemed as one of the most critical problems in three-dimensional (3D) massive multiple-input multiple-output (MIMO), which is recognized as the leading technology that enables 3D spatial signal processing in the fifth-generation (5G) wireless communications and beyond. Recently, by exploring the angular channel model and tensor decompositions, the accuracy of single-user channel estimation for 3D massive MIMO communications has been significantly improved given a limited number of pilot signals. However, these existing approaches cannot be straightforwardly extended to the multi-user channel estimation task, where the base station (BS) aims at acquiring the channels of multiple users at the same time. The difficulty is that the coupling among multiple users' channels makes the channel estimation deviate from widely-used tensor decompositions. It gives a non-standard tensor decomposition format that has not been well tackled. To overcome this challenge, besides directly fitting the new tensor model for channel estimation to the wireless data via block coordinate descent (BCD) method, which is prone to the overfitting of noises or requires regularization parameter tuning, we further propose a novel tuning-free channel estimation algorithm that can automatically control the channel model complexity and thus effectively avoid the overfitting. Numerical results are presented to demonstrate the excellent performance of the proposed algorithm in terms of both estimation accuracy and overfitting avoidance.

arXiv:2104.06799

Two-Dimensional DOA Estimation for L-shaped Nested Array via Tensor Modeling

The problem of two-dimensional (2-D) direction-of-arrival (DOA) estimation for the L-shaped nested array is considered. Typically, the multi-dimensional structure of the received signal in co-array domain is ignored in the problem considered. Moreover, the cross term generated by the correlated signal and noise components degrades the 2-D DOA estimation performance seriously. To tackle these issues, an iterative 2-D DOA estimation approach based on tensor modeling is proposed. To develop such approach, a higher-order tensor is constructed, whose factor matrices contain the sources azimuth and elevation information. By exploiting the Vandermonde structure of the factor matrix, a computationally efficient tensor decomposition method is then developed to estimate the sources DOA information in each dimension independently. The pair-matching of the azimuth and elevation angles is conducted via the cross-correlation matrix (CCM) of the received signals. An iterative method is further designed to improve the DOA estimation performance. Specifically, the cross term is estimated and removed in the next step of such iterative procedure on the basis of the DOA estimates originated from the tensor decomposition in the previous step. As a consequence, the DOA estimation with better accuracy and higher resolution is obtained. The proposed iterative 2-D DOA estimation method for the L-shaped nested array can resolve more sources than the number of real elements, which is superior to conventional approaches. Simulation results validate the performance improvement of the proposed 2-D DOA estimation method as compared to existing state-of-the-art DOA estimation techniques for the L-shaped nested array.

arXiv:2104.05758

Towards Extremely Compact RNNs for Video Recognition with Fully Decomposed Hierarchical Tucker Structure

Recurrent Neural Networks (RNNs) have been widely used in sequence analysis and modeling. However, when processing high-dimensional data, RNNs typically require very large model sizes, thereby bringing a series of deployment challenges. Although various prior works have been proposed to reduce the RNN model sizes, executing RNN models in resource-restricted environments is still a very challenging problem. In this paper, we propose to develop extremely compact RNN models with fully decomposed hierarchical Tucker (FDHT) structure. The HT decomposition does not only provide much higher storage cost reduction than the other tensor decomposition approaches but also brings better accuracy performance improvement for the compact RNN models. Meanwhile, unlike the existing tensor decomposition-based methods that can only decompose the input-to-hidden layer of RNNs, our proposed fully decomposition approach enables the comprehensive compression for the entire RNN models with maintaining very high accuracy. Our experimental results on several popular video recognition datasets show that our proposed fully decomposed hierarchical tucker-based LSTM (FDHT-LSTM) is extremely compact and highly efficient. To the best of our knowledge, FDHT-LSTM, for the first time, consistently achieves very high accuracy with only few thousand parameters (3,132 to 8,808) on different datasets. Compared with the state-of-the-art compressed RNN models, such as TT-LSTM, TR-LSTM and BT-LSTM, our FDHT-LSTM simultaneously enjoys both order-of-magnitude (3,985x to 10,711x) fewer parameters and significant accuracy improvement (0.6% to 12.7%).

arXiv:2104.05018공식 코드 · Papers with Code

TedNet: A Pytorch Toolkit for Tensor Decomposition Networks

Tensor Decomposition Networks (TDNs) prevail for their inherent compact architectures. To give more researchers a flexible way to exploit TDNs, we present a Pytorch toolkit named TedNet. TedNet implements 5 kinds of tensor decomposition(i.e., CANDECOMP/PARAFAC (CP), Block-Term Tucker (BTT), Tucker-2, Tensor Train (TT) and Tensor Ring (TR) on traditional deep neural layers, the convolutional layer and the fully-connected layer. By utilizing the basic layers, it is simple to construct a variety of TDNs. TedNet is available at https://github.com/tnbar/tednet.

논문 ID: document/8594834

Probabilistic Streaming Tensor Decomposition

Tensor decomposition is a fundamental tool for multiway data analysis. While most decomposition algorithms operate a collection of static data and perform batch processes, many applications produce data in a streaming manner - every time a subset of entries are generated, and previously seen entries cannot be revisited. In such scenarios, traditional decomposition approaches will be inappropriate, because they cannot provide timely updates when new data come in, and they need to access the whole dataset many times for batch optimization. To address this issue, we propose POST, a PrObabilistic Streaming Tensor decomposition algorithm, which enables real-time updates and predictions upon receiving new tensor entries, and supports dynamic growth of all the modes. Compared with the state-of-the-art streaming decomposition approach MAST, POST is more flexible in that it can handle arbitrary orders of streaming entries, and hence is more widely applicable. In addition, as a Bayesian inference algorithm, POST can quantify the uncertainty of the latent embeddings via their posterior distributions, and the confidence levels of the missing entry value predictions. On several real-world datasets, POST exhibits better or comparable predictive performance than MAST and other static decomposition algorithms.

arXiv:2103.11864

Recovery of Joint Probability Distribution from one-way marginals: Low rank Tensors and Random Projections

Joint probability mass function (PMF) estimation is a fundamental machine learning problem. The number of free parameters scales exponentially with respect to the number of random variables. Hence, most work on nonparametric PMF estimation is based on some structural assumptions such as clique factorization adopted by probabilistic graphical models, imposition of low rank on the joint probability tensor and reconstruction from 3-way or 2-way marginals, etc. In the present work, we link random projections of data to the problem of PMF estimation using ideas from tomography. We integrate this idea with the idea of low-rank tensor decomposition to show that we can estimate the joint density from just one-way marginals in a transformed space. We provide a novel algorithm for recovering factors of the tensor from one-way marginals, test it across a variety of synthetic and real-world datasets, and also perform MAP inference on the estimated model for classification.

arXiv:2103.06126

Spatial-Temporal Tensor Graph Convolutional Network for Traffic Prediction

Accurate traffic prediction is crucial to the guidance and management of urban traffics. However, most of the existing traffic prediction models do not consider the computational burden and memory space when they capture spatial-temporal dependence among traffic data. In this work, we propose a factorized Spatial-Temporal Tensor Graph Convolutional Network to deal with traffic speed prediction. Traffic networks are modeled and unified into a graph that integrates spatial and temporal information simultaneously. We further extend graph convolution into tensor space and propose a tensor graph convolution network to extract more discriminating features from spatial-temporal graph data. To reduce the computational burden, we take Tucker tensor decomposition and derive factorized a tensor convolution, which performs separate filtering in small-scale space, time, and feature modes. Besides, we can benefit from noise suppression of traffic data when discarding those trivial components in the process of tensor decomposition. Extensive experiments on two real-world traffic speed datasets demonstrate our method is more effective than those traditional traffic prediction methods, and meantime achieves state-of-the-art performance.

arXiv:2103.06234

Symmetry Breaking in Symmetric Tensor Decomposition

In this note, we consider the highly nonconvex optimization problem associated with computing the rank decomposition of symmetric tensors. We formulate the invariance properties of the loss function and show that critical points detected by standard gradient based methods are \emph{symmetry breaking} with respect to the target tensor. The phenomena, seen for different choices of target tensors and norms, make possible the use of recently developed analytic and algebraic tools for studying nonconvex optimization landscapes exhibiting symmetry breaking phenomena of similar nature.

arXiv:2103.02898공식 코드 · Papers with Code

Fast Tucker Rank Reduction for Non-Negative Tensors Using Mean-Field Approximation

We present an efficient low-rank approximation algorithm for non-negative tensors. The algorithm is derived from our two findings: First, we show that rank-1 approximation for tensors can be viewed as a mean-field approximation by treating each tensor as a probability distribution. Second, we theoretically provide a sufficient condition for distribution parameters to reduce Tucker ranks of tensors; interestingly, this sufficient condition can be achieved by iterative application of the mean-field approximation. Since the mean-field approximation is always given as a closed formula, our findings lead to a fast low-rank approximation algorithm without using a gradient method. We empirically demonstrate that our algorithm is faster than the existing non-negative Tucker rank reduction methods and achieves competitive or better approximation of given tensors.

arXiv:2103.00331공식 코드 · Papers with Code

CP-MDP: A CANDECOMP-PARAFAC Decomposition Approach to Solve a Markov Decision Process Multidimensional Problem

Markov Decision Process (MDP) is the underlying model for optimal planning for decision-theoretic agents in stochastic environments. Although much research focuses on solving MDP problems both in tabular form or using factored representations, none focused on tensor decomposition methods. Solving MDPs using tensor algebra offers the prospect of leveraging advances in tensor-based computations to further increase solver efficiency. In this paper, we develop an MDP solver for a multidimensional problem using a tensor decomposition method to compress the transition models and optimize the value iteration and policy iteration algorithms. We empirically evaluate our approach against tabular methods and show our approach can compute much larger problems using substantially less memory, opening up new possibilities for tensor-based approaches in stochastic planning