LID 2020: The Learning from Imperfect Data Challenge Results

Learning from imperfect data becomes an issue in many industrial applications after the research community has made profound progress in supervised learning from perfectly annotated datasets. The purpose of the Learning from Imperfect Data (LID) workshop is to inspire and facilitate the research in developing novel approaches that would harness the imperfect data and improve the data-efficiency during training. A massive amount of user-generated data nowadays available on multiple internet services. How to leverage those and improve the machine learning models is a high impact problem. We organize the challenges in conjunction with the workshop. The goal of these challenges is to find the state-of-the-art approaches in the weakly supervised learning setting for object detection, semantic segmentation, and scene parsing. There are three tracks in the challenge, i.e., weakly supervised semantic segmentation (Track 1), weakly supervised scene parsing (Track 2), and weakly supervised object localization (Track 3). In Track 1, based on ILSVRC DET, we provide pixel-level annotations of 15K images from 200 categories for evaluation. In Track 2, we provide point-based annotations for the training set of ADE20K. In Track 3, based on ILSVRC CLS-LOC, we provide pixel-level annotations of 44,271 images for evaluation. Besides, we further introduce a new evaluation metric proposed by \cite{zhang2020rethinking}, i.e., IoU curve, to measure the quality of the generated object localization maps. This technical report summarizes the highlights from the challenge. The challenge submission server and the leaderboard will continue to open for the researchers who are interested in it. More details regarding the challenge and the benchmarks are available at https://lidchallenge.github.io

Collaborative Video Object Segmentation by Multi-Scale Foreground-Background Integration

This paper investigates the principles of embedding learning to tackle the challenging semi-supervised video object segmentation. Unlike previous practices that focus on exploring the embedding learning of foreground object (s), we consider background should be equally treated. Thus, we propose a Collaborative video object segmentation by Foreground-Background Integration (CFBI) approach. CFBI separates the feature embedding into the foreground object region and its corresponding background region, implicitly promoting them to be more contrastive and improving the segmentation results accordingly. Moreover, CFBI performs both pixel-level matching processes and instance-level attention mechanisms between the reference and the predicted sequence, making CFBI robust to various object scales. Based on CFBI, we introduce a multi-scale matching structure and propose an Atrous Matching strategy, resulting in a more robust and efficient framework, CFBI+. We conduct extensive experiments on two popular benchmarks, i.e., DAVIS and YouTube-VOS. Without applying any simulated data for pre-training, our CFBI+ achieves the performance (J&F) of 82.9% and 82.8%, outperforming all the other state-of-the-art methods. Code: https://github.com/z-x-yang/CFBI.

Simple and Effective Few-Shot Named Entity Recognition with Structured Nearest Neighbor Learning

We present a simple few-shot named entity recognition (NER) system based on nearest neighbor learning and structured inference. Our system uses a supervised NER model trained on the source domain, as a feature extractor. Across several test domains, we show that a nearest neighbor classifier in this feature-space is far more effective than the standard meta-learning approaches. We further propose a cheap but effective method to capture the label dependencies between entity tags without expensive CRF training. We show that our method of combining structured decoding with nearest neighbor learning achieves state-of-the-art performance on standard few-shot NER evaluation tasks, improving F1 scores by $6\%$ to $16\%$ absolute points over prior meta-learning based systems.

DOTS: Decoupling Operation and Topology in Differentiable Architecture Search

Differentiable Architecture Search (DARTS) has attracted extensive attention due to its efficiency in searching for cell structures. However, DARTS mainly focuses on the operation search, leaving the cell topology implicitly depending on the searched operation weights. Hence, a problem is raised: can cell topology be well represented by the operation weights? The answer is negative because we observe that the operation weights fail to indicate the performance of cell topology. In this paper, we propose to Decouple the Operation and Topology Search (DOTS), which decouples the cell topology representation from the operation weights to make an explicit topology search. DOTS is achieved by defining an additional cell topology search space besides the original operation search space. Within the DOTS framework, we propose group annealing operation search and edge annealing topology search to bridge the optimization gap between the searched over-parameterized network and the derived child network. DOTS is efficient and only costs 0.2 and 1 GPU-day to search the state-of-the-art cell architectures on CIFAR and ImageNet, respectively. By further searching for the topology of DARTS' searched cell, we can improve DARTS' performance significantly. The code will be publicly available.

Tasks Integrated Networks: Joint Detection and Retrieval for Image Search

The traditional object retrieval task aims to learn a discriminative feature representation with intra-similarity and inter-dissimilarity, which supposes that the objects in an image are manually or automatically pre-cropped exactly. However, in many real-world searching scenarios (e.g., video surveillance), the objects (e.g., persons, vehicles, etc.) are seldom accurately detected or annotated. Therefore, object-level retrieval becomes intractable without bounding-box annotation, which leads to a new but challenging topic, i.e. image-level search. In this paper, to address the image search issue, we first introduce an end-to-end Integrated Net (I-Net), which has three merits: 1) A Siamese architecture and an on-line pairing strategy for similar and dissimilar objects in the given images are designed. 2) A novel on-line pairing (OLP) loss is introduced with a dynamic feature dictionary, which alleviates the multi-task training stagnation problem, by automatically generating a number of negative pairs to restrict the positives. 3) A hard example priority (HEP) based softmax loss is proposed to improve the robustness of classification task by selecting hard categories. With the philosophy of divide and conquer, we further propose an improved I-Net, called DC-I-Net, which makes two new contributions: 1) two modules are tailored to handle different tasks separately in the integrated framework, such that the task specification is guaranteed. 2) A class-center guided HEP loss (C2HEP) by exploiting the stored class centers is proposed, such that the intra-similarity and inter-dissimilarity can be captured for ultimate retrieval. Extensive experiments on famous image-level search oriented benchmark datasets demonstrate that the proposed DC-I-Net outperforms the state-of-the-art tasks-integrated and tasks-separated image search models.

Each Part Matters: Local Patterns Facilitate Cross-view Geo-localization

Cross-view geo-localization is to spot images of the same geographic target from different platforms, e.g., drone-view cameras and satellites. It is challenging in the large visual appearance changes caused by extreme viewpoint variations. Existing methods usually concentrate on mining the fine-grained feature of the geographic target in the image center, but underestimate the contextual information in neighbor areas. In this work, we argue that neighbor areas can be leveraged as auxiliary information, enriching discriminative clues for geo-localization. Specifically, we introduce a simple and effective deep neural network, called Local Pattern Network (LPN), to take advantage of contextual information in an end-to-end manner. Without using extra part estimators, LPN adopts a square-ring feature partition strategy, which provides the attention according to the distance to the image center. It eases the part matching and enables the part-wise representation learning. Owing to the square-ring partition design, the proposed LPN has good scalability to rotation variations and achieves competitive results on two prevailing benchmarks, i.e., University-1652 and CVUSA. Besides, we also show the proposed LPN can be easily embedded into other frameworks to further boost performance.

Point Adversarial Self Mining: A Simple Method for Facial Expression Recognition in the Wild

In this paper, the Point Adversarial Self Mining (PASM) approach, a simple yet effective way to progressively mine knowledge from training samples, is proposed to produce training data for CNNs to improve the performance and network generality in Facial Expression Recognition (FER) task. In order to achieve a high prediction accuracy under real-world scenarios, most of the existing works choose to manipulate the network architectures and design sophisticated loss terms. Although demonstrated to be effective in real scenarios, those aforementioned methods require extra efforts in network design. Inspired by random erasing and adversarial erasing, we propose PASM for data augmentation, simulating the data distribution in the wild. Specifically, given a sample and a pre-trained network, our proposed approach locates the informative region in the sample generated by point adversarial attack policy. The informative region is highly structured and sparse. Comparing to the regions produced by random erasing which selects the region in a purely random way and adversarial erasing which operates by attention maps, the located informative regions obtained by PASM are more adaptive and better aligned with the previous findings: not all but only a few facial regions contribute to the accurate prediction. Then, the located informative regions are masked out from the original samples to generate augmented images, which would force the network to explore additional information from other less informative regions. The augmented images are used to finetune the network to enhance its generality. In the refinement process, we take advantage of knowledge distillation, utilizing the pre-trained network to provide guidance and retain knowledge from old samples to train a new network with the same structural configuration.

Describing Unseen Videos via Multi-Modal Cooperative Dialog Agents

With the arising concerns for the AI systems provided with direct access to abundant sensitive information, researchers seek to develop more reliable AI with implicit information sources. To this end, in this paper, we introduce a new task called video description via two multi-modal cooperative dialog agents, whose ultimate goal is for one conversational agent to describe an unseen video based on the dialog and two static frames. Specifically, one of the intelligent agents - Q-BOT - is given two static frames from the beginning and the end of the video, as well as a finite number of opportunities to ask relevant natural language questions before describing the unseen video. A-BOT, the other agent who has already seen the entire video, assists Q-BOT to accomplish the goal by providing answers to those questions. We propose a QA-Cooperative Network with a dynamic dialog history update learning mechanism to transfer knowledge from A-BOT to Q-BOT, thus helping Q-BOT to better describe the video. Extensive experiments demonstrate that Q-BOT can effectively learn to describe an unseen video by the proposed model and the cooperative learning method, achieving the promising performance where Q-BOT is given the full ground truth history dialog.

DONet: Dual Objective Networks for Skin Lesion Segmentation

Skin lesion segmentation is a crucial step in the computer-aided diagnosis of dermoscopic images. In the last few years, deep learning based semantic segmentation methods have significantly advanced the skin lesion segmentation results. However, the current performance is still unsatisfactory due to some challenging factors such as large variety of lesion scale and ambiguous difference between lesion region and background. In this paper, we propose a simple yet effective framework, named Dual Objective Networks (DONet), to improve the skin lesion segmentation. Our DONet adopts two symmetric decoders to produce different predictions for approaching different objectives. Concretely, the two objectives are actually defined by different loss functions. In this way, the two decoders are encouraged to produce differentiated probability maps to match different optimization targets, resulting in complementary predictions accordingly. The complementary information learned by these two objectives are further aggregated together to make the final prediction, by which the uncertainty existing in segmentation maps can be significantly alleviated. Besides, to address the challenge of large variety of lesion scales and shapes in dermoscopic images, we additionally propose a recurrent context encoding module (RCEM) to model the complex correlation among skin lesions, where the features with different scale contexts are efficiently integrated to form a more robust representation. Extensive experiments on two popular benchmarks well demonstrate the effectiveness of the proposed DONet. In particular, our DONet achieves 0.881 and 0.931 dice score on ISIC 2018 and $\text{PH}^2$, respectively. Code will be made public available.

Describing Unseen Videos via Multi-ModalCooperative Dialog Agents

With the arising concerns for the AI systems provided with direct access to abundant sensitive information, researchers seek to develop more reliable AI with implicit information sources. To this end, in this paper, we introduce a new task called video description via two multi-modal cooperative dialog agents, whose ultimate goal is for one conversational agent to describe an unseen video based on the dialog and two static frames. Specifically, one of the intelligent agents - Q-BOT - is given two static frames from the beginning and the end of the video, as well as a finite number of opportunities to ask relevant natural language questions before describing the unseen video. A-BOT, the other agent who has already seen the entire video, assists Q-BOT to accomplish the goal by providing answers to those questions. We propose a QA-Cooperative Network with a dynamic dialog history update learning mechanism to transfer knowledge from A-BOT to Q-BOT, thus helping Q-BOT to better describe the video. Extensive experiments demonstrate that Q-BOT can effectively learn to describe an unseen video by the proposed model and the cooperative learning method, achieving the promising performance where Q-BOT is given the full ground truth history dialog.