Convolutional neural networks (CNNs) learn to extract representations of complex features, such as object shapes and textures to solve image recognition tasks. Recent work indicates that CNNs trained on ImageNet are biased towards features that encode textures and that these alone are sufficient to generalize to unseen test data from the same distribution as the training data but often fail to generalize to out-of-distribution data. It has been shown that augmenting the training data with different image styles decreases this texture bias in favor of increased shape bias while at the same time improving robustness to common corruptions, such as noise and blur. Commonly, this is interpreted as shape bias increasing corruption robustness. However, this relationship is only hypothesized. We perform a systematic study of different ways of composing inputs based on natural images, explicit edge information, and stylization. While stylization is essential for achieving high corruption robustness, we do not find a clear correlation between shape bias and robustness. We conclude that the data augmentation caused by style-variation accounts for the improved corruption robustness and increased shape bias is only a byproduct.
Overtaking is one of the most challenging tasks in driving, and the current solutions to autonomous overtaking are limited to simple and static scenarios. In this paper, we present a method for behaviour and trajectory planning for safe autonomous overtaking. The proposed method optimizes the trajectory by simultaneously enforcing safety and minimizing intrusion onto the adjacent lane. Furthermore, the method allows the overtaking to be aborted, enabling the autonomous vehicle to merge back in the lane, if safety is compromised, because of e.g. traffic in opposing direction appearing during the maneuver execution. A finite state machine is used to select an appropriate maneuver at each time, and a combination of safe and reachable sets is used to iteratively generate intermediate reference targets based on the current maneuver. A nonlinear model predictive controller then plans dynamically feasible and collision-free trajectories to these intermediate reference targets. Simulation experiments demonstrate that the combination of intermediate reference generation and model predictive control is able to handle multiple behaviors, including following a lead vehicle, overtaking and aborting the overtake, within a single framework.
High-stakes applications require AI-generated models to be interpretable. Current algorithms for the synthesis of potentially interpretable models rely on objectives or regularization terms that represent interpretability only coarsely (e.g., model size) and are not designed for a specific user. Yet, interpretability is intrinsically subjective. In this paper, we propose an approach for the synthesis of models that are tailored to the user by enabling the user to steer the model synthesis process according to her or his preferences. We use a bi-objective evolutionary algorithm to synthesize models with trade-offs between accuracy and a user-specific notion of interpretability. The latter is estimated by a neural network that is trained concurrently to the evolution using the feedback of the user, which is collected using uncertainty-based active learning. To maximize usability, the user is only asked to tell, given two models at the time, which one is less complex. With experiments on two real-world datasets involving 61 participants, we find that our approach is capable of learning estimations of interpretability that can be very different for different users. Moreover, the users tend to prefer models found using the proposed approach over models found using non-personalized interpretability indices.
Can we distinguish between two wireless transmitters sending exactly the same message, using the same protocol? The opportunity for doing so arises due to subtle nonlinear variations across transmitters, even those made by the same manufacturer. Since these effects are difficult to model explicitly, we investigate learning device fingerprints using complex-valued deep neural networks (DNNs) that take as input the complex baseband signal at the receiver. Such fingerprints should be robust to ID spoofing, and to distribution shifts across days and locations due to clock drift and variations in the wireless channel. In this paper, we point out that, unless proactively discouraged from doing so, DNNs learn these strong confounding features rather than the subtle nonlinear characteristics that are the basis for stable signatures. Thus, a network trained on data collected during one day performs poorly on a different day, and networks allowed access to post-preamble information rely on easily-spoofed ID fields. We propose and evaluate strategies, based on augmentation and estimation, to promote generalization across realizations of these confounding factors, using data from WiFi and ADS-B protocols. We conclude that, while DNN training has the advantage of not requiring explicit signal models, significant modeling insights are required to focus the learning on the effects we wish to capture.
This paper presents a biologically plausible method for converting real-valued input into spike trains for processing with spiking neural networks. The proposed method mimics the adaptive behaviour of retinal ganglion cells and allows input neurons to adapt their response to changes in the statistics of the input. Thus, rather than passively receiving values and forwarding them to the hidden and output layers, the input layer acts as a self-regulating filter which emphasises deviations from the average while allowing the input neurons to become effectively desensitised to the average itself. Another merit of the proposed method is that it requires only one input neuron per variable, rather than an entire population of neurons as in the case of the commonly used conversion method based on Gaussian receptive fields. In addition, since the statistics of the input emerge naturally over time, it becomes unnecessary to pre-process the data before feeding it to the network. This enables spiking neural networks to process raw, non-normalised streaming data. A proof-of-concept experiment is performed to demonstrate that the proposed method operates as expected.
Getting the distance to objects is crucial for autonomous vehicles. In instances where depth sensors cannot be used, this distance has to be estimated from RGB cameras. As opposed to cars, the task of estimating depth from on-board mounted cameras is made complex on drones because of the lack of constrains on motion during flights. %In the case of drones, this task is even more complex than for car-mounted cameras since the camera motion is unconstrained. In this paper, we present a method to estimate the distance of objects seen by an on-board mounted camera by using its RGB video stream and drone motion information. Our method is built upon a pyramidal convolutional neural network architecture and uses time recurrence in pair with geometric constraints imposed by motion to produce pixel-wise depth maps. %from a RGB video stream of a camera attached to the drone In our architecture, each level of the pyramid is designed to produce its own depth estimate based on past observations and information provided by the previous level in the pyramid. We introduce a spatial reprojection layer to maintain the spatio-temporal consistency of the data between the levels. We analyse the performance of our approach on Mid-Air, a public drone dataset featuring synthetic drone trajectories recorded in a wide variety of unstructured outdoor environments. Our experiments show that our network outperforms state-of-the-art depth estimation methods and that the use of motion information is the main contributing factor for this improvement. The code of our method is publicly available on GitHub; see $\href{https://github.com/michael-fonder/M4Depth}{\text{https://github.com/michael-fonder/M4Depth}}$
Discrete event systems (DES) have been established and deeply developed in the framework of probabilistic and fuzzy computing models due to the necessity of practical applications in fuzzy and probabilistic systems. With the development of quantum computing and quantum control, a natural problem is to simulate DES by means of quantum computing models and to establish {\it quantum DES} (QDES). The motivation is twofold: on the one hand, QDES have potential applications when DES are simulated and processed by quantum computers, where quantum systems are employed to simulate the evolution of states driven by discrete events, and on the other hand, QDES may have essential advantages over DES concerning state complexity for imitating some practical problems. The goal of this paper is to establish a basic framework of QDES by using {\it quantum finite automata} (QFA) as the modelling formalisms, and the supervisory control theorems of QDES are established and proved. Then we present a polynomial-time algorithm to decide whether or not the controllability condition holds. In particular, we construct a number of new examples of QFA to illustrate the supervisory control of QDES and to verify the essential advantages of QDES over DES in state complexity.
Motivated by the recent interest in cyber-physical and autonomous robotic systems, we study the problem of dynamically coupled multi-agent systems under a set of signal temporal logic tasks. In particular, the satisfaction of each of these signal temporal logic tasks depends on the behavior of a distinct set of agents. Instead of abstracting the agent dynamics and the temporal logic tasks into a discrete domain and solving the problem therein or using optimization-based methods, we derive collaborative feedback control laws. These control laws are based on a decentralized control barrier function condition that results in discontinuous control laws, as opposed to a centralized condition resembling the single-agent case. The benefits of our approach are inherent robustness properties typically present in feedback control as well as satisfaction guarantees for continuous-time multi-agent systems. More specifically, time-varying control barrier functions are used that account for the semantics of the signal temporal logic tasks at hand. For a certain fragment of signal temporal logic tasks, we further propose a systematic way to construct such control barrier functions. Finally, we show the efficacy and robustness of our framework in an experiment including a group of three omnidirectional robots.
We present a synthetic prediction market whose agent purchase logic is defined using a sigmoid transformation of a convex semi-algebraic set defined in feature space. Asset prices are determined by a logarithmic scoring market rule. Time varying asset prices affect the structure of the semi-algebraic sets leading to time-varying agent purchase rules. We show that under certain assumptions on the underlying geometry, the resulting synthetic prediction market can be used to arbitrarily closely approximate a binary function defined on a set of input data. We also provide sufficient conditions for market convergence and show that under certain instances markets can exhibit limit cycles in asset spot price. We provide an evolutionary algorithm for training agent parameters to allow a market to model the distribution of a given data set and illustrate the market approximation using two open source data sets. Results are compared to standard machine learning methods.
We present a new trainable system for physically plausible markerless 3D human motion capture, which achieves state-of-the-art results in a broad range of challenging scenarios. Unlike most neural methods for human motion capture, our approach, which we dub physionical, is aware of physical and environmental constraints. It combines in a fully differentiable way several key innovations, i.e., 1. a proportional-derivative controller, with gains predicted by a neural network, that reduces delays even in the presence of fast motions, 2. an explicit rigid body dynamics model and 3. a novel optimisation layer that prevents physically implausible foot-floor penetration as a hard constraint. The inputs to our system are 2D joint keypoints, which are canonicalised in a novel way so as to reduce the dependency on intrinsic camera parameters -- both at train and test time. This enables more accurate global translation estimation without generalisability loss. Our model can be finetuned only with 2D annotations when the 3D annotations are not available. It produces smooth and physically principled 3D motions in an interactive frame rate in a wide variety of challenging scenes, including newly recorded ones. Its advantages are especially noticeable on in-the-wild sequences that significantly differ from common 3D pose estimation benchmarks such as Human 3.6M and MPI-INF-3DHP. Qualitative results are available at http://gvv.mpi-inf.mpg.de/projects/PhysAware/