Vanilla models for object detection and instance segmentation suffer from the heavy bias toward detecting frequent objects in the long-tailed setting. Existing methods address this issue mostly during training, e.g., by re-sampling or re-weighting. In this paper, we investigate a largely overlooked approach -- post-processing calibration of confidence scores. We propose NorCal, Normalized Calibration for long-tailed object detection and instance segmentation, a simple and straightforward recipe that reweighs the predicted scores of each class by its training sample size. We show that separately handling the background class and normalizing the scores over classes for each proposal are keys to achieving superior performance. On the LVIS dataset, NorCal can effectively improve nearly all the baseline models not only on rare classes but also on common and frequent classes. Finally, we conduct extensive analysis and ablation studies to offer insights into various modeling choices and mechanisms of our approach.

Convolutional neural networks (CNNs) often have poor generalization performance under domain shift. One way to improve domain generalization is to collect diverse source data from multiple relevant domains so that a CNN model is allowed to learn more domain-invariant, and hence generalizable representations. In this work, we address domain generalization with MixStyle, a plug-and-play, parameter-free module that is simply inserted to shallow CNN layers and requires no modification to training objectives. Specifically, MixStyle probabilistically mixes feature statistics between instances. This idea is inspired by the observation that visual domains can often be characterized by image styles which are in turn encapsulated within instance-level feature statistics in shallow CNN layers. Therefore, inserting MixStyle modules in effect synthesizes novel domains albeit in an implicit way. MixStyle is not only simple and flexible, but also versatile -- it can be used for problems whereby unlabeled images are available, such as semi-supervised domain generalization and unsupervised domain adaptation, with a simple extension to mix feature statistics between labeled and pseudo-labeled instances. We demonstrate through extensive experiments that MixStyle can significantly boost the out-of-distribution generalization performance across a wide range of tasks including object recognition, instance retrieval, and reinforcement learning.

Applying neural network (NN) methods in games can lead to various With the recent boom in neural network (NN) applications, game new and exciting game dynamics not previously possible. However, designers have been increasingly exploring a variety of NN approaches they also lead to new challenges such as the lack of large, in computer games [65]. These include approaches where clean datasets, varying player skill levels, and changing gameplay a NN is directly incorporated into the gameplay experience or as a strategies. In this paper, we focus on the adversarial player strategy method for dynamically generating content that would otherwise aspect in the game iNNk, in which players try to communicate be created by a human artist [65]. This approach has been utilized secret code words through drawings with the goal of not being in well-known games such as Black and White [36], Creatures [23], deciphered by a NN. Some strategies exploit weaknesses in the NN and Forza Motosport [17], which adapt game agent behavior in response that consistently trick it into making incorrect classifications, leading to player input. In these cases, the NN makes gameplay to unbalanced gameplay. We present a method that combines more personalized and potentially more engaging.

Although recent works have developed methods that can generate estimations (or imputations) of the missing entries in a dataset to facilitate downstream analysis, most depend on assumptions that may not align with real-world applications and could suffer from poor performance in subsequent tasks. This is particularly true if the data have large missingness rates or a small population. More importantly, the imputation error could be propagated into the prediction step that follows, causing the gradients used to train the prediction models to be biased. Consequently, in this work, we introduce the importance guided stochastic gradient descent (IGSGD) method to train multilayer perceptrons (MLPs) and long short-term memories (LSTMs) to directly perform inference from inputs containing missing values without imputation. Specifically, we employ reinforcement learning (RL) to adjust the gradients used to train the models via back-propagation. This not only reduces bias but allows the model to exploit the underlying information behind missingness patterns. We test the proposed approach on real-world time-series (i.e., MIMIC-III), tabular data obtained from an eye clinic, and a standard dataset (i.e., MNIST), where our imputation-free predictions outperform the traditional two-step imputation-based predictions using state-of-the-art imputation methods.

The last decade has seen a significant increase of interest in deep learning research, with many public successes that have demonstrated its potential. As such, these systems are now being incorporated into commercial products. With this comes an additional challenge: how can we build AI systems that solve tasks where there is not a crisp, well-defined specification? While multiple solutions have been proposed, in this competition we focus on one in particular: learning from human feedback. Rather than training AI systems using a predefined reward function or using a labeled dataset with a predefined set of categories, we instead train the AI system using a learning signal derived from some form of human feedback, which can evolve over time as the understanding of the task changes, or as the capabilities of the AI system improve. The MineRL BASALT competition aims to spur forward research on this important class of techniques. We design a suite of four tasks in Minecraft for which we expect it will be hard to write down hardcoded reward functions. These tasks are defined by a paragraph of natural language: for example, "create a waterfall and take a scenic picture of it", with additional clarifying details. Participants must train a separate agent for each task, using any method they want. Agents are then evaluated by humans who have read the task description. To help participants get started, we provide a dataset of human demonstrations on each of the four tasks, as well as an imitation learning baseline that leverages these demonstrations. Our hope is that this competition will improve our ability to build AI systems that do what their designers intend them to do, even when the intent cannot be easily formalized. Besides allowing AI to solve more tasks, this can also enable more effective regulation of AI systems, as well as making progress on the value alignment problem.

Advances in AI, and especially machine learning, are increasingly drawing research interest and efforts towards predictive process monitoring, the subfield of process mining (PM) that concerns predicting next events, process outcomes and remaining execution times. Unfortunately, researchers use a variety of datasets and ways to split them into training and test sets. The documentation of these preprocessing steps is not always complete. Consequently, research results are hard or even impossible to reproduce and to compare between papers. At times, the use of non-public domain knowledge further hampers the fair competition of ideas. Often the training and test sets are not completely separated, a data leakage problem particular to predictive process monitoring. Moreover, test sets usually suffer from bias in terms of both the mix of case durations and the number of running cases. These obstacles pose a challenge to the field's progress. The contribution of this paper is to identify and demonstrate the importance of these obstacles and to propose preprocessing steps to arrive at unbiased benchmark datasets in a principled way, thus creating representative test sets without data leakage with the aim of levelling the playing field, promoting open science and contributing to more rapid progress in predictive process monitoring.

Person re-identification via 3D skeletons is an emerging topic with great potential in security-critical applications. Existing methods typically learn body and motion features from the body-joint trajectory, whereas they lack a systematic way to model body structure and underlying relations of body components beyond the scale of body joints. In this paper, we for the first time propose a Self-supervised Multi-scale Skeleton Graph Encoding (SM-SGE) framework that comprehensively models human body, component relations, and skeleton dynamics from unlabeled skeleton graphs of various scales to learn an effective skeleton representation for person Re-ID. Specifically, we first devise multi-scale skeleton graphs with coarse-to-fine human body partitions, which enables us to model body structure and skeleton dynamics at multiple levels. Second, to mine inherent correlations between body components in skeletal motion, we propose a multi-scale graph relation network to learn structural relations between adjacent body-component nodes and collaborative relations among nodes of different scales, so as to capture more discriminative skeleton graph features. Last, we propose a novel multi-scale skeleton reconstruction mechanism to enable our framework to encode skeleton dynamics and high-level semantics from unlabeled skeleton graphs, which encourages learning a discriminative skeleton representation for person Re-ID. Extensive experiments show that SM-SGE outperforms most state-of-the-art skeleton-based methods. We further demonstrate its effectiveness on 3D skeleton data estimated from large-scale RGB videos. Our codes are open at https://github.com/Kali-Hac/SM-SGE.

Rap generation, which aims to produce lyrics and corresponding singing beats, needs to model both rhymes and rhythms. Previous works for rap generation focused on rhyming lyrics but ignored rhythmic beats, which are important for rap performance. In this paper, we develop DeepRapper, a Transformer-based rap generation system that can model both rhymes and rhythms. Since there is no available rap dataset with rhythmic beats, we develop a data mining pipeline to collect a large-scale rap dataset, which includes a large number of rap songs with aligned lyrics and rhythmic beats. Second, we design a Transformer-based autoregressive language model which carefully models rhymes and rhythms. Specifically, we generate lyrics in the reverse order with rhyme representation and constraint for rhyme enhancement and insert a beat symbol into lyrics for rhythm/beat modeling. To our knowledge, DeepRapper is the first system to generate rap with both rhymes and rhythms. Both objective and subjective evaluations demonstrate that DeepRapper generates creative and high-quality raps with rhymes and rhythms. Code will be released on GitHub.

Deployed machine learning models are confronted with the problem of changing data over time, a phenomenon also called concept drift. While existing approaches of concept drift detection already show convincing results, they require true labels as a prerequisite for successful drift detection. Especially in many real-world application scenarios--like the ones covered in this work--true labels are scarce, and their acquisition is expensive. Therefore, we introduce a new algorithm for drift detection, Uncertainty Drift Detection (UDD), which is able to detect drifts without access to true labels. Our approach is based on the uncertainty estimates provided by a deep neural network in combination with Monte Carlo Dropout. Structural changes over time are detected by applying the ADWIN technique on the uncertainty estimates, and detected drifts trigger a retraining of the prediction model. In contrast to input data-based drift detection, our approach considers the effects of the current input data on the properties of the prediction model rather than detecting change on the input data only (which can lead to unnecessary retrainings). We show that UDD outperforms other state-of-the-art strategies on two synthetic as well as ten real-world data sets for both regression and classification tasks.

ABSTRACT We explore the potential to control terrain vehicles using deep reinforcement in scenarios where human operators and traditional control methods are inadequate. This letter presents a controller that perceives, plans, and successfully controls a 16-tonne forestry vehicle with two frame articulation joints, six wheels, and their actively articulated suspensions to traverse rough terrain. The carefully shaped reward signal promotes safe, environmental, and efficient driving, which leads to the emergence of unprecedented driving skills. We test learned skills in a virtual environment, including terrains reconstructed from high-density laser scans of forest sites. The results confirm that deep reinforcement learning has the potential to enhance control of vehicles with complex dynamics and high-dimensional observation data compared to human operators or traditional control methods, especially in rough terrain. 1 INTRODUCTION Deep reinforcement learning has recently shown promise for locomotion tasks, but its usefulness to learn control of heavy vehicles in rough terrain is widely unknown. Conventionally, the design of rough terrain vehicles strives to promote high traversability and be easily operated by humans. The drivelines involve differentials and bogie suspension that provide ground compliance and reduces the many degrees of freedom, leaving only speed and heading for the operator to control. An attractive alternative is to use actively articulated suspensions and individual wheel control. These have the potential to reduce the energy consumption and ground damage, yet increase traversability and tip over stability [11, 6, 21, 10, 9].