We study the training dynamics of shallow neural networks, in a two-timescale regime in which the stepsizes for the inner layer are much smaller than those for the outer layer. In this regime, we prove convergence of the gradient flow to a global optimum of the non-convex optimization problem in a simple univariate setting. The number of neurons need not be asymptotically large for our result to hold, distinguishing our result from popular recent approaches such as the neural tangent kernel or mean-field regimes. Experimental illustration is provided, showing that the stochastic gradient descent behaves according to our description of the gradient flow and thus converges to a global optimum in the two-timescale regime, but can fail outside of this regime.
相關內容
神經網絡（Neural Networks）是世界上三個最古老的神經建模學會的檔案期刊:國際神經網絡學會(INNS)、歐洲神經網絡學會(ENNS)和日本神經網絡學會(JNNS)。神經網絡提供了一個論壇，以發展和培育一個國際社會的學者和實踐者感興趣的所有方面的神經網絡和相關方法的計算智能。神經網絡歡迎高質量論文的提交，有助于全面的神經網絡研究，從行為和大腦建模，學習算法，通過數學和計算分析，系統的工程和技術應用，大量使用神經網絡的概念和技術。這一獨特而廣泛的范圍促進了生物和技術研究之間的思想交流，并有助于促進對生物啟發的計算智能感興趣的跨學科社區的發展。因此，神經網絡編委會代表的專家領域包括心理學，神經生物學，計算機科學，工程，數學，物理。該雜志發表文章、信件和評論以及給編輯的信件、社論、時事、軟件調查和專利信息。文章發表在五個部分之一:認知科學，神經科學，學習系統，數學和計算分析、工程和應用。
官網地址：
This study explores the robustness of label noise classifiers, aiming to enhance model resilience against noisy data in complex real-world scenarios. Label noise in supervised learning, characterized by erroneous or imprecise labels, significantly impairs model performance. This research focuses on the increasingly pertinent issue of label noise's impact on practical applications. Addressing the prevalent challenge of inaccurate training data labels, we integrate adversarial machine learning (AML) and importance reweighting techniques. Our approach involves employing convolutional neural networks (CNN) as the foundational model, with an emphasis on parameter adjustment for individual training samples. This strategy is designed to heighten the model's focus on samples critically influencing performance.
The impact of outliers and anomalies on model estimation and data processing is of paramount importance, as evidenced by the extensive body of research spanning various fields over several decades: thousands of research papers have been published on the subject. As a consequence, numerous reviews, surveys, and textbooks have sought to summarize the existing literature, encompassing a wide range of methods from both the statistical and data mining communities. While these endeavors to organize and summarize the research are invaluable, they face inherent challenges due to the pervasive nature of outliers and anomalies in all data-intensive applications, irrespective of the specific application field or scientific discipline. As a result, the resulting collection of papers remains voluminous and somewhat heterogeneous. To address the need for knowledge organization in this domain, this paper implements the first systematic meta-survey of general surveys and reviews on outlier and anomaly detection. Employing a classical systematic survey approach, the study collects nearly 500 papers using two specialized scientific search engines. From this comprehensive collection, a subset of 56 papers that claim to be general surveys on outlier detection is selected using a snowball search technique to enhance field coverage. A meticulous quality assessment phase further refines the selection to a subset of 25 high-quality general surveys. Using this curated collection, the paper investigates the evolution of the outlier detection field over a 20-year period, revealing emerging themes and methods. Furthermore, an analysis of the surveys sheds light on the survey writing practices adopted by scholars from different communities who have contributed to this field. Finally, the paper delves into several topics where consensus has emerged from the literature. These include taxonomies of outlier types, challenges posed by high-dimensional data, the importance of anomaly scores, the impact of learning conditions, difficulties in benchmarking, and the significance of neural networks. Non-consensual aspects are also discussed, particularly the distinction between local and global outliers and the challenges in organizing detection methods into meaningful taxonomies.
We devise, implement and performance-asses DYAD, a layer which can serve as a faster and more memory-efficient approximate replacement for linear layers, (nn.Linear() in Pytorch). These layers appear in common subcomponents, such as in the ff module of Transformers. DYAD is based on a bespoke near-sparse matrix structure which approximates the dense "weight" matrix W that matrix-multiplies the input in the typical realization of such a layer, a.k.a DENSE. Our alternative near-sparse matrix structure is decomposable to a sum of 2 matrices permutable to a block-sparse counterpart. These can be represented as 3D tensors, which in unison allow a faster execution of matrix multiplication with the mini-batched input matrix X compared to DENSE (O(rows(W ) x cols(W )) --> O( rows(W ) x cols(W ) # of blocks )). As the crux of our experiments, we pretrain both DYAD and DENSE variants of 2 sizes of the OPT arch and 1 size of the Pythia arch, including at different token scales of the babyLM benchmark. We find DYAD to be competitive (>= 90%) of DENSE performance on zero-shot (e.g. BLIMP), few-shot (OPENLM) and finetuning (GLUE) benchmarks, while being >=7-15% faster to train on-GPU even at 125m scale, besides surfacing larger speedups at increasing scale and model width.
In several branches of the social sciences and humanities, surveys based on standardized questionnaires are a prominent research tool. While there are a variety of ways to analyze the data, some standard procedures have become established. When those surveys want to analyze differences in the answer patterns of different groups (e.g., countries, gender, age, ...), these procedures can only be carried out in a meaningful way if there is measurement invariance, i.e., the measured construct has psychometric equivalence across groups. As recently raised as an open problem by Sauerwein et al. (2021), new evaluation methods that work in the absence of measurement invariance are needed. This paper promotes an unsupervised learning-based approach to such research data by proposing a procedure that works in three phases: data preparation, clustering of questionnaires, and measuring similarity based on the obtained clustering and the properties of each group. We generate synthetic data in three data sets, which allows us to compare our approach with the PCA approach under measurement invariance and under violated measurement invariance. As a main result, we obtain that the approach provides a natural comparison between groups and a natural description of the response patterns of the groups. Moreover, it can be safely applied to a wide variety of data sets, even in the absence of measurement invariance. Finally, this approach allows us to translate (violations of) measurement invariance into a meaningful measure of similarity.
We adopt the integral definition of the fractional Laplace operator and study an optimal control problem on Lipschitz domains that involves a fractional elliptic partial differential equation (PDE) as state equation and a control variable that enters the state equation as a coefficient; pointwise constraints on the control variable are considered as well. We establish the existence of optimal solutions and analyze first and, necessary and sufficient, second order optimality conditions. Regularity estimates for optimal variables are also analyzed. We develop two finite element discretization strategies: a semidiscrete scheme in which the control variable is not discretized, and a fully discrete scheme in which the control variable is discretized with piecewise constant functions. For both schemes, we analyze the convergence properties of discretizations and derive error estimates.
Due to the dynamic characteristics of instantaneity and steepness, employing domain decomposition techniques for simulating rogue wave solutions is highly appropriate. Wherein, the backward compatible PINN (bc-PINN) is a temporally sequential scheme to solve PDEs over successive time segments while satisfying all previously obtained solutions. In this work, we propose improvements to the original bc-PINN algorithm in two aspects based on the characteristics of error propagation. One is to modify the loss term for ensuring backward compatibility by selecting the earliest learned solution for each sub-domain as pseudo reference solution. The other is to adopt the concatenation of solutions obtained from individual subnetworks as the final form of the predicted solution. The improved backward compatible PINN (Ibc-PINN) is applied to study data-driven higher-order rogue waves for the nonlinear Schr\"{o}dinger (NLS) equation and the AB system to demonstrate the effectiveness and advantages. Transfer learning and initial condition guided learning (ICGL) techniques are also utilized to accelerate the training. Moreover, the error analysis is conducted on each sub-domain and it turns out that the slowdown of Ibc-PINN in error accumulation speed can yield greater advantages in accuracy. In short, numerical results fully indicate that Ibc-PINN significantly outperforms bc-PINN in terms of accuracy and stability without sacrificing efficiency.
This study elaborates a text-based metric to quantify the unique position of stylized scientific research, characterized by its innovative integration of diverse knowledge components and potential to pivot established scientific paradigms. Our analysis reveals a concerning decline in stylized research, highlighted by its comparative undervaluation in terms of citation counts and protracted peer-review duration. Despite facing these challenges, the disruptive potential of stylized research remains robust, consistently introducing groundbreaking questions and theories. This paper posits that substantive reforms are necessary to incentivize and recognize the value of stylized research, including optimizations to the peer-review process and the criteria for evaluating scientific impact. Embracing these changes may be imperative to halt the downturn in stylized research and ensure enduring scholarly exploration in endless frontiers.
We present an approach for analyzing message passing graph neural networks (MPNNs) based on an extension of graphon analysis to a so called graphon-signal analysis. A MPNN is a function that takes a graph and a signal on the graph (a graph-signal) and returns some value. Since the input space of MPNNs is non-Euclidean, i.e., graphs can be of any size and topology, properties such as generalization are less well understood for MPNNs than for Euclidean neural networks. We claim that one important missing ingredient in past work is a meaningful notion of graph-signal similarity measure, that endows the space of inputs to MPNNs with a regular structure. We present such a similarity measure, called the graphon-signal cut distance, which makes the space of all graph-signals a dense subset of a compact metric space -- the graphon-signal space. Informally, two deterministic graph-signals are close in cut distance if they ``look like'' they were sampled from the same random graph-signal model. Hence, our cut distance is a natural notion of graph-signal similarity, which allows comparing any pair of graph-signals of any size and topology. We prove that MPNNs are Lipschitz continuous functions over the graphon-signal metric space. We then give two applications of this result: 1) a generalization bound for MPNNs, and, 2) the stability of MPNNs to subsampling of graph-signals. Our results apply to any regular enough MPNN on any distribution of graph-signals, making the analysis rather universal.
We hypothesize that due to the greedy nature of learning in multi-modal deep neural networks, these models tend to rely on just one modality while under-fitting the other modalities. Such behavior is counter-intuitive and hurts the models' generalization, as we observe empirically. To estimate the model's dependence on each modality, we compute the gain on the accuracy when the model has access to it in addition to another modality. We refer to this gain as the conditional utilization rate. In the experiments, we consistently observe an imbalance in conditional utilization rates between modalities, across multiple tasks and architectures. Since conditional utilization rate cannot be computed efficiently during training, we introduce a proxy for it based on the pace at which the model learns from each modality, which we refer to as the conditional learning speed. We propose an algorithm to balance the conditional learning speeds between modalities during training and demonstrate that it indeed addresses the issue of greedy learning. The proposed algorithm improves the model's generalization on three datasets: Colored MNIST, Princeton ModelNet40, and NVIDIA Dynamic Hand Gesture.
Artificial neural networks thrive in solving the classification problem for a particular rigid task, acquiring knowledge through generalized learning behaviour from a distinct training phase. The resulting network resembles a static entity of knowledge, with endeavours to extend this knowledge without targeting the original task resulting in a catastrophic forgetting. Continual learning shifts this paradigm towards networks that can continually accumulate knowledge over different tasks without the need to retrain from scratch. We focus on task incremental classification, where tasks arrive sequentially and are delineated by clear boundaries. Our main contributions concern 1) a taxonomy and extensive overview of the state-of-the-art, 2) a novel framework to continually determine the stability-plasticity trade-off of the continual learner, 3) a comprehensive experimental comparison of 11 state-of-the-art continual learning methods and 4 baselines. We empirically scrutinize method strengths and weaknesses on three benchmarks, considering Tiny Imagenet and large-scale unbalanced iNaturalist and a sequence of recognition datasets. We study the influence of model capacity, weight decay and dropout regularization, and the order in which the tasks are presented, and qualitatively compare methods in terms of required memory, computation time, and storage.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191