Understanding generalization in deep learning is arguably one of the most important questions in deep learning. Deep learning has been successfully adopted to a large number of problems ranging from pattern recognition to complex decision making, but many recent researchers have raised many concerns about deep learning, among which the most important is generalization. Despite numerous attempts, conventional statistical learning approaches have yet been able to provide a satisfactory explanation on why deep learning works. A recent line of works aims to address the problem by trying to predict the generalization performance through complexity measures. In this competition, we invite the community to propose complexity measures that can accurately predict generalization of models. A robust and general complexity measure would potentially lead to a better understanding of deep learning's underlying mechanism and behavior of deep models on unseen data, or shed light on better generalization bounds. All these outcomes will be important for making deep learning more robust and reliable. * Lead organizer: Yiding Jiang; Scott Yak and Pierre Foret help implement large portion of the infrastructure and the remaining organizers' order is randomized.
We empirically show that the test error of deep networks can be estimated by simply training the same architecture on the same training set but with a different run of Stochastic Gradient Descent (SGD), and measuring the disagreement rate between the two networks on unlabeled test data. This builds onand is a stronger version of -the observation in Nakkiran & Bansal (2020), which requires the second run to be on an altogether fresh training set. We further theoretically show that this peculiar phenomenon arises from the well-calibrated nature of ensembles of SGD-trained models. This finding not only provides a simple empirical measure to directly predict the test error using unlabeled test data, but also establishes a new conceptual connection between generalization and calibration.
No abstract
A major component of overfitting in model-free reinforcement learning (RL) involves the case where the agent may mistakenly correlate reward with certain spurious features from the observations generated by the Markov Decision Process (MDP). We provide a general framework for analyzing this scenario, which we use to design multiple synthetic benchmarks from only modifying the observation space of an MDP. When an agent overfits to different observation spaces even if the underlying MDP dynamics is fixed, we term this observational overfitting. Our experiments expose intriguing properties especially with regards to implicit regularization, and also corroborate results from previous works in RL generalization and supervised learning (SL). * Work partially performed as an OpenAI Fellow. † Work performed during the Google AI Residency Program.
As a result of the trend toward economic globalization, the vigorous development of cross-border e-commerce has attracted many scholars to study this field, involving many related fields, such as consumer behavior, advertising, information systems, and supply chain management. Throughout the existing literature, it can be found that most of the research focuses on certain influencing factors of the development of cross-border e-commerce, and there is no systematic and macro- overview of the development trend of research in this field in recent years, nor the integration and analysis of keywords. Given that the research in the field of cross-border e-commerce is fragmented to a large extent, to effectively explore the research trend in this field, we must understand the current situation of cross-border e-commerce. Systematic bibliometric analysis can solve this problem by providing publishing trends and information on various topics. Therefore, based on the scientific database web, this study collected 198 references related to cross-border e-commerce from 2016 to 2021, briefly introduced the current situation and development trend of cross-border e-commerce, visually analyzed and refined the journals, authors, research institutions, countries, publication years, keywords, citations of academic publications in this field, and other key information, and summarized the development trend and path of CEBC in the existing research. It is helpful for researchers to solve the current research gap, understand the future research direction in this field, and help academia establish a strict knowledge system.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.