Nanjing University

About: Nanjing University is a education organization based out in Nanjing, China. It is known for research contribution in the topics: Catalysis & Population. The organization has 85961 authors who have published 105504 publications receiving 2289036 citations. The organization is also known as: NJU & Nanking University.

BBN: Bilateral-Branch Network with Cumulative Learning for Long-Tailed Visual Recognition

Abstract: Our work focuses on tackling the challenging but natural visual recognition task of long-tailed data distribution (i.e., a few classes occupy most of the data, while most classes have rarely few samples). In the literature, class re-balancing strategies (e.g., re-weighting and re-sampling) are the prominent and effective methods proposed to alleviate the extreme imbalance for dealing with long-tailed problems. In this paper, we firstly discover that these re-balancing methods achieving satisfactory recognition accuracy owe to that they could significantly promote the classifier learning of deep networks. However, at the same time, they will unexpectedly damage the representative ability of the learned deep features to some extent. Therefore, we propose a unified Bilateral-Branch Network (BBN) to take care of both representation learning and classifier learning simultaneously, where each branch does perform its own duty separately. In particular, our BBN model is further equipped with a novel cumulative learning strategy, which is designed to first learn the universal patterns and then pay attention to the tail data gradually. Extensive experiments on four benchmark datasets, including the large-scale iNaturalist ones, justify that the proposed BBN can significantly outperform state-of-the-art methods. Furthermore, validation experiments can demonstrate both our preliminary discovery and effectiveness of tailored designs in BBN for long-tailed problems. Our method won the first place in the iNaturalist 2019 large scale species classification competition, and our code is open-source and available at this https URL.

A Class of Projection and Contraction Methods for Monotone Variational-Inequalities

Abstract: In this paper we introduce a new class of iterative methods for solving the monotone variational inequalities $$u* \in \Omega , (u - u*)^T F(u*) \geqslant 0, \forall u \in \Omega $$ Each iteration of the methods presented consists essentially only of the computation ofF(u), a projection to Ω,v:=P Ω[u-F(u)], and the mappingF(v) The distance of the iterates to the solution set monotonically converges to zero Both the methods and the convergence proof are quite simple

China's Natural Wetlands: Past Problems, Current Status, and Future Challenges

Abstract: Natural wetlands, occupying 3.8% of China's land and providing 54.9% of ecosystem services, are unevenly distributed among eight wetland regions. Natural wetlands in China suffered great loss and degradation (e.g., 23.0% freshwater swamps, 51.2% costal wetlands) because of the wetland reclamation during China's long history of civilization, and the population pressure and the misguided policies over the last 50 years. Recently, with an improved understanding that healthy wetland ecosystems play a vital role in her sustainable economic development, China started major efforts in wetland conservation, as signified by the policy to return reclaimed croplands to wetlands, the funding of billions of dollars to restore degraded wetlands, and the national plan to place 90% of natural wetlands under protection by 2030. This paper describes the current status of the natural wetlands in China, reviews past problems, and discusses current efforts and future challenges in protecting China's natural wetlands.

Deep Forest

Abstract: Current deep learning models are mostly build upon neural networks, i.e., multiple layers of parameterized differentiable nonlinear modules that can be trained by backpropagation. In this paper, we explore the possibility of building deep models based on non-differentiable modules. We conjecture that the mystery behind the success of deep neural networks owes much to three characteristics, i.e., layer-by-layer processing, in-model feature transformation and sufficient model complexity. We propose the gcForest approach, which generates \textit{deep forest} holding these characteristics. This is a decision tree ensemble approach, with much less hyper-parameters than deep neural networks, and its model complexity can be automatically determined in a data-dependent way. Experiments show that its performance is quite robust to hyper-parameter settings, such that in most cases, even across different data from different domains, it is able to get excellent performance by using the same default setting. This study opens the door of deep learning based on non-differentiable modules, and exhibits the possibility of constructing deep models without using backpropagation.