Intelligent tutoring systems (ITS) typically employ scaffolding to provide learning opportunities for needed students. To assess the ITS effectiveness, most past research relies on the end-of-year exam outcome. However, it is important to assess ITS' immediate effectiveness after the scaffolding. To address this gap, this paper draws principles from theory on scaffolding and zone of proximal development to guide the analysis of micro-level, time-stamped student action data from the ASSISTments. Using a deep neural network to analyze the data, we reveal that the appropriate use of scaffolds in the ITS fosters modest growth of math problem-solving skills. Excessive use, however, is counterproductive when students are overly reliant on ITS. Our findings suggest that scaffolding in ITS or AI tutoring should encourage student active learning and prevent over-reliance on intelligent systems.

Standard deep reinforcement learning (DRL) aims to maximize expected reward, considering collected experiences equally in formulating a policy. This differs from human decision-making, where gains and losses are valued differently and outlying outcomes are given increased consideration. It also fails to capitalize on opportunities to improve safety and/or performance through the incorporation of distributional context. Several approaches to distributional DRL have been investigated, with one popular strategy being to evaluate the projected distribution of returns for possible actions. We propose a more direct approach whereby risk-sensitive objectives, specified in terms of the cumulative distribution function (CDF) of the distribution of full-episode rewards, are optimized. This approach allows for outcomes to be weighed based on relative quality, can be used for both continuous and discrete action spaces, and may naturally be applied in both constrained and unconstrained settings. We show how to compute an asymptotically consistent estimate of the policy gradient for a broad class of risk-sensitive objectives via sampling, subsequently incorporating variance reduction and regularization measures to facilitate effective on-policy learning. We then demonstrate that the use of moderately "pessimistic" risk profiles, which emphasize scenarios where the agent performs poorly, leads to enhanced exploration and a continual focus on addressing deficiencies. We test the approach using different risk profiles in six OpenAI Safety Gym environments, comparing to state of the art on-policy methods. Without cost constraints, we find that pessimistic risk profiles can be used to reduce cost while improving total reward accumulation. With cost constraints, they are seen to provide higher positive rewards than risk-neutral approaches at the prescribed allowable cost.

In contrast to perturbation-based attacks, patch-based attacks are physically realizable, and are therefore increasingly studied. However, prior work neglects the possibility of adaptive attacks optimized for 3D pose. For the first time, to our knowledge, we consider the challenge of designing and evaluating attacks on image sequences using 3D optimization along entire 3D kinematic trajectories. In this context, we study a type of dynamic attack, referred to as "adaptive active attacks" (AAA), that takes into consideration the pose of the observer being targeted. To better address the threat and risk posed by AAA attacks, we develop several novel risk-based and trajectory-based metrics. These are designed to capture the risk of attack success for attacking earlier in the trajectory to derail autonomous driving systems as well as tradeoffs that may arise given the possibility of additional detection. We evaluate performance of white-box targeted attacks using a subset of ImageNet classes, and demonstrate, in aggregate, that AAA attacks can pose threats beyond static attacks in kinematic settings in situations of predominantly looming motion (i. e., a prevalent use case in automated vehicular navigation). Results demonstrate that AAA attacks can exhibit targeted attack success exceeding 10% in aggregate, and for some specific classes, up to 15% over their static counterparts. However, taking into consideration the probability of detection by the defender shows a more nuanced risk pattern. These new insights are important for guiding future adversarial machine learning studies and suggest researchers should consider defense against novel threats posed by dynamic attacks for full trajectories and videos.

Despite the progress in deep learning networks, efficient learning at the edge (enabling adaptable, low-complexity machine learning solutions) remains a critical need for defense and commercial applications. We have pursued multiple neuroscience-inspired AI efforts which may overcome these data inefficiencies, power inefficiencies, and lack of generalization. We envision a pipeline to utilize large neuroimaging datasets, including maps of the brain which capture neuron and synapse connectivity, to improve machine learning approaches. We have pursued different approaches within this pipeline structure, including data-driven discovery in biological networks, augmenting existing computational neuroscience models, investigating the biological structure which enables behaviors, and modifying existing machine learning architectures with insight from biological structure. First, as a demonstration of data-driven discovery, the team has developed a technique for discovery of repeated subcircuits, or motifs. These were incorporated into a neural architecture search approach to evolve network architectures. Second, we have conducted analysis of the heading direction circuit in the fruit fly, which performs fusion of visual and angular velocity features, to explore augmenting existing computational models with new insight. Our team discovered a novel pattern of connectivity, implemented a new model, and demonstrated sensor fusion on a robotic platform. Third, the team analyzed circuitry for memory formation in the fruit fly connectome, enabling the design of a novel generative replay approach. This replay approach resulted in an over 20% accuracy improvement in an incremental class learning scenario, and also demonstrated the ability to utilize large neuroscience datasets to analyze the neural connectivity underlying behavior. Finally, the team has begun analysis of connectivity in mammalian cortex to explore potential improvements to transformer networks. These constraints increased network robustness on the most challenging examples in the CIFAR-10-C computer vision robustness benchmark task, while reducing learnable attention parameters by over an order of magnitude. Taken together, these results demonstrate multiple potential approaches to utilize insight from neural systems for developing robust and efficient machine learning techniques.

Standard deep reinforcement learning (DRL) agents aim to maximize expected reward, considering collected experiences equally in formulating a policy. This differs from human decision-making, where gains and losses are valued differently and outlying outcomes are given increased consideration. It also wastes an opportunity for the agent to modulate behavior based on distributional context. Several approaches to distributional DRL have been investigated, with one popular strategy being to evaluate the projected distribution of returns for possible actions. We propose a more direct approach, whereby the distribution of full-episode outcomes is optimized to maximize a chosen function of its cumulative distribution function (CDF). This technique allows for outcomes to be weighed based on relative quality, does not require modification of the reward function to modulate agent behavior, and may be used for both continuous and discrete action spaces. We show how to achieve an unbiased estimate of the policy gradient for a broad class of CDF-based objectives via sampling, subsequently incorporating variance reduction measures to facilitate effective on-policy learning. We use the resulting approach to train agents with different “risk profiles” in penalty-based formulations of six OpenAI Safety Gym environments, finding that moderate emphasis on improvement in training scenarios where the agent performs poorly generally improves agent behavior. We interpret and explore this observation, which leads to improved performance over the widely-used Proximal Policy Optimization algorithm in all environments tested.

Human-aware robot navigation promises a range of applications in which mobile robots bring versatile assistance to people in common human environments. While prior research has mostly focused on modeling pedestrians as independent, intentional individuals, people move in groups; consequently, it is imperative for mobile robots to respect human groups when navigating around people. This paper explores learning group-aware navigation policies based on dynamic group formation using deep reinforcement learning. Through simulation experiments, we show that group-aware policies, compared to baseline policies that neglect human groups, achieve greater robot navigation performance (e.g., fewer collisions), minimize violation of social norms and discomfort, and reduce the robot's movement impact on pedestrians. Our results contribute to the development of social navigation and the integration of mobile robots into human environments.

This paper explores learning group-aware navigation policies based on dynamic group formation using deep reinforcement learning. Through simulation experiments, we show that group-aware policies achieve greater robot navigation performance (e.g., fewer collisions), minimize violation of social norms and discomfort, and reduce the robot’s movement impact on pedestrians when compared to the baseline.

Current deep reinforcement learning (DRL) methods fail to address risk in an intelligent manner, potentially leading to unsafe behaviors when deployed. One strategy for improving agent risk management is to mimic human behavior. While imperfect, human risk processing displays two key benefits absent from standard artificial agents: accounting for rare but consequential events and incorporating context. The for- mer ability may prevent catastrophic outcomes in unfamil-iar settings while the latter results in asymmetric process- ing of potential gains and losses. These two attributes have been quantified by behavioral economists and form the ba- sis of cumulative prospect theory (CPT), a leading model of human decision-making. We introduce a two-step method for training DRL agents to maximize the CPT-value of full-episode rewards accumulated from an environment, rather than the standard practice of maximizing expected discounted rewards. We quantitatively compare the distribution of out- comes when optimizing full-episode expected reward, CPT-value, and conditional value-at-risk (CVaR) in the CrowdSim robot navigation environment, elucidating the impacts of different objectives on the agent’s willingness to trade safety for speed. We find that properly-configured maximization of CPT-value allows for a reduction of the frequency of negative outcomes with only a slight degradation of the best outcomes, compared to maximization of expected reward.

Recent breakthroughs and rapid progress in AI will impact, if not transform, every mission. JHU/APL developed an AI Technology Roadmap to guide the Laboratory’s contributions to the critical challenges the nation will face developing and implementing intelligent systems for these missions over the coming decades. We began this exercise by describing a series of envisioned futures for intelligent systems across sea, land, air, space and information, and examined them to identify the common AI technology vectors needed to achieve each vision: (1) Autonomous Perception, describing the path to intelligent systems that perceive in the context of the extreme uncertainty and complexity of the real world; (2) Superhuman Decision-Making and Autonomous Action, to realize the potential for intelligent systems to reason over more information than any team of analysts or operators and act in ways systems under manned control cannot; (3) Human-Machine Teaming at the Speed of Thought, to ensure humans can stay involved at speed and scale; and of particular importance for national security applications, (4) Safe and Assured Operation, so these systems can be trusted to stay true to commander’s intent, in adversarial and sensitive contexts. Each technology vector is aligned with a targeted goal, and with each goal we provide a roadmap in the form of near-, mid-, and long-term AI advances critical to reaching the goal. This paper describes the JHU/APL AI Technology Roadmap and presents key examples of recent progress and forward-looking research and exploratory development along each vector.

Accurate state and uncertainty estimation is imperative for mobile robots and self driving vehicles to achieve safe navigation in pedestrian rich environments. A critical component of state and uncertainty estimation for robot navigation is to perform robustly under out-of-distribution noise. Traditional methods of state estimation decouple perception and state estimation making it difficult to operate on noisy, high dimensional data. Here, we describe an approach that combines the expressiveness of deep neural networks with principled approaches to uncertainty estimation found in recursive filters. We particularly focus on techniques that provide better robustness to out-of-distribution noise and demonstrate applicability of our approach on two scenarios: a simple noisy pendulum state estimation problem and real world pedestrian localization using the nuScenes dataset. We show that our approach improves state and uncertainty estimation compared to baselines while achieving approximately 3x improvement in computational efficiency.

We focus on the development of effective adversarial patch attacks and – for the first time – jointly address the antagonistic objectives of attack success and obtrusiveness via the design of novel semi-transparent patches. This work is motivated by our pursuit of a systematic performance analysis of patch attack robustness with regard to geometric transformations. Specifically, we first elucidate a) key factors underpinning patch attack success and b) the impact of distributional shift between training and testing/deployment when cast under the Expectation over Transformation (EoT) formalism. By focusing our analysis on three principal classes of transformations (rotation, scale, and location), our findings provide quantifiable insights into the design of effective patch attacks and demonstrate that scale, among all factors, significantly impacts patch attack success. Working from these findings, we then focus on addressing how to overcome the principal limitations of scale for the deployment of attacks in real physical settings: namely the obtrusiveness of large patches. Our strategy is to turn to the novel design of irregularly-shaped, semi-transparent partial patches which we construct via a new optimization process that jointly addresses the antagonistic goals of mitigating obtrusiveness and maximizing effectiveness. Our study – we hope – will help encourage more focus in the community on the issues of obtrusiveness, scale, and success in patch attacks.

With deep neural networks (DNNs) being used increasingly in many applications, it is critical to improve our understanding of their failure modes and potential mitigations. A Johns Hopkins University Applied Physics Laboratory (APL) team successfully inserted a backdoor (train-time attack) into a common object detection model. In conjunction with this research, they developed a principled methodology to evaluate patch attacks (test-time attacks) and the factors impacting their success. Their approach enabled the creation of a novel optimization framework for the first-ever design of semitransparent patches that can overcome scale limitations while retaining desirable factors with regard to deployment and detectability.

Mobile robots capable of navigating seamlessly and safely in pedestrian rich environments promise to bring robotic assistance closer to our daily lives. In this paper we draw on insights of how humans move in crowded spaces to explore how to recognize pedestrian navigation intent, how to predict pedestrian motion and how a robot may adapt its navigation policy dynamically when facing unexpected human movements. Our approach is to develop algorithms that replicate this behavior. We experimentally demonstrate the effectiveness of our prediction algorithm using real-world pedestrian datasets and achieve comparable or better prediction accuracy compared to several state-of-the-art approaches. Moreover, we show that confidence of pedestrian prediction can be used to adjust the risk of a navigation policy adaptively to afford the most comfortable level as measured by the frequency of personal space violation in comparison with baselines. Furthermore, our adaptive navigation policy is able to reduce the number of collisions by 43% in the presence of novel pedestrian motion not seen during training.

Our work focuses on unsupervised and generative methods that address the following goals: (1) learning unsupervised generative representations that discover latent factors controlling image semantic attributes, (2) studying how this ability to control attributes formally relates to the issue of latent factor disentanglement, clarifying related but dissimilar concepts that had been confounded in the past, and (3) developing anomaly detection methods that leverage representations learned in the first goal. For goal 1, we propose a network architecture that exploits the combination of multiscale generative models with mutual information (MI) maximization. For goal 2, we derive an analytical result, lemma 1, that brings clarity to two related but distinct concepts: the ability of generative networks to control semantic attributes of images they generate, resulting from MI maximization, and the ability to disentangle latent space representations, obtained via total correlation minimization. More specifically, we demonstrate that maximizing semantic attribute control encourages disentanglement of latent factors. Using lemma 1 and adopting MI in our loss function, we then show empirically that for image generation tasks, the proposed approach exhibits superior performance as measured in the quality and disentanglement of the generated images when compared to other state-of-the-art methods, with quality assessed via the Fréchet inception distance (FID) and disentanglement via mutual information gap. For goal 3, we design several systems for anomaly detection exploiting representations learned in goal 1 and demonstrate their performance benefits when compared to state-of-the-art generative and discriminative algorithms. Our contributions in representation learning have potential applications in addressing other important problems in computer vision, such as bias and privacy in AI.

Human-aware robot navigation promises a range of applications in which mobile robots bring versatile assistance to people in common human environments. While prior research has mostly focused on modeling pedestrians as independent, intentional individuals, people move in groups; consequently, it is imperative for mobile robots to respect human groups when navigating around people. This paper explores learning group-aware navigation policies based on dynamic group formation using deep reinforcement learning. Through simulation experiments, we show that group-aware policies, compared to baseline policies that neglect human groups, achieve greater robot navigation performance (e.g., fewer collisions), minimize violation of social norms and discomfort, and reduce the robot’s movement impact on pedestrians. Our results contribute to the development of social navigation and the integration of mobile robots into human environments.

Target localization in WSNs (Wireless Sensor Networks) has been an active research area due to its importance and wide range of applications. Most of the proposals so far, however, have focused on cooperative targets that in one way or another work with the localization infrastructure. For example, mobile targets can emit RF or acoustic signals either on purpose (ActiveBadges) or by their nature (e.g., bird chirps). Instead, we focus on localizing non-cooperative targets using active sensors. Specifically, we use a network of mono-static low-power, pulsed, Dopper radars that independently estimate the target's Doppler velocity with respect to each radar. These measurements are then used to estimate the target's location and true velocity using an Extended Kalman filter. We have built a prototype of this system using TelosB motes and BumbleBee radars.

Fast, collision-free motion through human environments remains a challenging problem for robotic systems. In these situations, the robot’s ability to reason about its future motion and other agents is often severely limited. By contrast, biological systems routinely make decisions by taking into consideration what might exist in the future based on prior experience. In this paper, we present an approach that provides robotic systems the ability to make future predictions of the environment. We evaluate several deep network architectures, including purely generative and adversarial models for map prediction. We further extend this approach to predict future pedestrian motion. We show that prediction plays a key role in enabling an adaptive, risk-sensitive control policy. Our algorithms are able to generate future maps with a structural similarity index metric up to 0.899 compared to the ground truth map. Further, our adaptive crowd navigation algorithm is able to reduce the number of collisions by 43% in the presence of novel pedestrian motion not seen during training.

The ability to create artificial intelligence (AI) capable of performing complex tasks is rapidly outpacing our ability to ensure the safe and assured operation of AI-enabled systems. Fortunately, a landscape of AI safety research is emerging in response to this asymmetry and yet there is a long way to go. In particular, recent simulation environments created to illustrate AI safety risks are relatively simple or narrowly-focused on a particular issue. Hence, we see a critical need for AI safety research environments that abstract essential aspects of complex real-world applications. In this work, we introduce the AI safety TanksWorld as an environment for AI safety research with three essential aspects: competing performance objectives, human-machine teaming, and multi-agent competition. The AI safety TanksWorld aims to accelerate the advancement of safe multi-agent decision-making algorithms by providing a software framework to support competitions with both system performance and safety objectives. As a work in progress, this paper introduces our research objectives and learning environment with reference code and baseline performance metrics to follow in a future work.

A critical capability required for the wide adoption of mobile robots into society is the ability to navigate safely around pedestrians. One important component to enable safe navigation is to accurately predict the motion of pedestrians in the scene. The main objective of this research is to develop novel techniques that accurately predict human motion by using past motion and intent as a prior for making the prediction. In this study, we develop neural network architectures that are capable of learning environment-agnostic embeddings that serve as a prior for prediction. We combine these embeddings with contextual information including desired velocity and a probability distribution describing the intent to make predictions. We compare the average displacement error and final displacement error with state-of-the-art published results and show evidence that combining contextual information results in more accurate prediction of future motion.