Out-of-Distribution (OOD) detection is a crucial problem for the safe deployment of machine learning models identifying samples that fall outside of the training distribution, i.e. in-distribution data (ID). Most OOD works focus on the classification models trained with Cross Entropy (CE) and attempt to fix its inherent issues. In this work we leverage powerful representation learned with Supervised Contrastive (SupCon) training and propose a holistic approach to learn a classifier robust to OOD data. We extend SupCon loss with two additional contrast terms. The first term pushes auxiliary OOD representations away from ID representations without imposing any constraints on similarities among auxiliary data. The second term pushes OOD features far from the existing class prototypes, while pushing ID representations closer to their corresponding class prototype. When auxiliary OOD data is not available, we propose feature mixing techniques to efficiently generate pseudo-OOD features. Our solution is simple and efficient and acts as a natural extension of the closed-set supervised contrastive representation learning. We compare against different OOD detection methods on the common benchmarks and show state-of-the-art results.

Cross entropy loss has served as the main objective function for classification-based tasks. Widely deployed for learning neural network classifiers, it shows both effectiveness and a probabilistic interpretation. Recently, after the success of self supervised contrastive representation learning methods, supervised contrastive methods have been proposed to learn representations and have shown superior and more robust performance, compared to solely training with cross entropy loss. However, cross entropy loss is still needed to train the final classification layer. In this work, we investigate the possibility of learning both the representation and the classifier using one objective function that combines the robustness of contrastive learning and the probabilistic interpretation of cross entropy loss. First, we revisit a previously proposed contrastive-based objective function that approximates cross entropy loss and present a simple extension to learn the classifier jointly. Second, we propose a new version of the supervised contrastive training that learns jointly the parameters of the classifier and the backbone of the network. We empirically show that these proposed objective functions demonstrate state-of-the-art performance and show a significant improvement over the standard cross entropy loss with more training stability and robustness in various challenging settings.

While unlabelled data can be largely available and even abundant, the annotation process can be quite expensive and limiting. Under the assumption that some samples are more important for a given task than others, active learning targets the problem of identifying the most informative samples that one should acquire annotations for. In this work we propose a simple sample selection criterion that moves beyond the conventional reliance on model uncertainty as proxy to leverage new labels. By first accepting the model prediction and then judging its effect on the generalization error, we can better identify wrongly predicted samples. We also present a very efficient approximation to our criterion, providing a similarity-based interpretation. In addition to evaluating our method on the standard benchmarks of active learning, we consider the challenging yet realistic imbalanced data scenario. We show state-of-the-art results, especially on the imbalanced setting, and achieve better rates at identifying wrongly predicted samples than existing active learning methods. Our method is simple, model agnostic and relies on the current model status without the need for re-training from scratch.

We present a novel approach to the detection of unknown objects in the context of autonomous driving. The problem is formulated as anomaly detection, since we assume that the unknown stuff or object appearance cannot be learned. To that end, we propose a reconstruction module that can be used with many existing semantic segmentation networks, and that is trained to recognize and reconstruct road (drivable) surface from a small bottleneck. We postulate that poor reconstruction of the road surface is due to areas that are outside of the training distribution, which is a strong indicator of an anomaly. The road structural similarity error is coupled with the semantic segmentation to incorporate information from known classes and produce final per-pixel anomaly scores. The proposed JSR-Net was evaluated on four datasets, Lost-and-found, Road Anomaly, Road Obstacles, and FishyScapes, achieving state-of-art performance on all, reducing the false positives significantly, while typically having the highest average precision for wide range of operation points.

In this paper, we introduce a game in which the player navigates an avatar through a maze by using a brain-computer interface (BCI) that analyzes the steady-state visual evoked potential (SSVEP) responses recorded with electroencephalography (EEG) on the player's scalp. The four-command control game, called The Maze, was specifically designed around an SSVEP BCI and validated in several EEG setups when using a traditional electrode cap with relocatable electrodes and a consumer-grade headset with fixed electrodes (Emotiv EPOC). We experimentally derive the parameter values that provide an acceptable tradeoff between accuracy of game control and interactivity, and evaluate the control provided by the BCI during gameplay. As a final step in the validation of the game, a population study on a broad audience was conducted with the EPOC headset in a real-world setting. The study revealed that the majority (85%) of the players enjoyed the game in spite of its intricate control (mean accuracy 80.37%, mean mission time ratio 0.90). We also discuss what to take into account while designing BCI-based games.

We propose a complex-valued multilayer feedforward neural network classifier for decoding of phase-coded information from steady-state visual evoked potentials. To optimize the performance of the classifier we supply it with two filter-based feature selection strategies. The proposed approaches could be used for a phase-coded brain-computer interface, enabling to encode several targets using only one stimulation frequency. The proposed classifier is a multichannel one, which distinguishes our approach from the existing single-channel ones. We show that the proposed approach outperforms others in terms of accuracy and length of the data segments used for decoding. We show that the decoding based on one optimally selected channel yields an inferior performance compared to the one based on several features, which supports our argument for a multichannel approach.

To offer increased security and comfort, advanced driver-assistance systems (ADASs) should consider individual driving styles. Here, we present a system that learns a human's basic driving behavior and demonstrate its use as ADAS by issuing alerts when detecting inconsistent driving behavior. In contrast to much other work in this area, which is based on or obtained from simulation, our system is implemented as a multithreaded parallel central processing unit (CPU)/graphics processing unit (GPU) architecture in a real car and trained with real driving data to generate steering and acceleration control for road following. It also implements a method for detecting independently moving objects (IMOs) for spotting obstacles. Both learning and IMO detection algorithms are data driven and thus improve above the limitations of model-based approaches. The system's ability to imitate the teacher's behavior is analyzed on known and unknown streets, and results suggest its use for steering assistance but limit the use of the acceleration signal to curve negotiation. We propose that this ability to adapt to the driver can lead to better acceptance of ADAS, which is an important sales argument.

We present two neural models for vergence angle control of a robotic head, a simplified and a more complex one. Both models work in a closed-loop manner and do not rely on explicitly computed disparity, but extract the desired vergence angle from the post-processed response of a population of disparity tuned complex cells, the actual gaze direction and the actual vergence angle. The first model assumes that the gaze direction of the robotic head is orthogonal to its baseline and the stimulus is a frontoparallel plane orthogonal to the gaze direction. The second model goes beyond these assumptions, and operates reliably in the general case where all restrictions on the orientation of the gaze, as well as the stimulus position, type and orientation, are dropped.

The Mind Speller is a Brain-Computer Interface which enables subjects to spell text on a computer screen by detecting P300 Event-Related Potentials in their electroencephalograms. This BCI application is of particular interest for disabled patients who have lost all means of verbal and motor communication. We report on the implementation of a feature extraction procedure on a new ultra low-power 8-channel wireless EEG device. The feature extraction procedure is based on downsampled EEG signal epochs, the Student's t-statistic of the Continuous Wavelet Transform, and the Common Spatial Pattern technique. For classification, we use a linear Least-Squares Support Vector Machine. The results show that subjects are potentially able to communicate a character in less than ten seconds with an accuracy of 94.5%, which is more than twice as fast as the state of the art. In addition since our EEG device is wireless it offers an increased comfort to the subject.