Search

Human comprehension of a video stream is naturally broad: in a few instants, we are able to understand what is happening, the relevance and relationship of objects, and forecast what will follow in the near future, everything all at once. We believe that - to effectively transfer such an holistic perception to intelligent machines - an important role is played by learning to correlate concepts and to abstract knowledge coming from different tasks, to synergistically exploit them when learning novel skills. To accomplish this, we seek for a unified approach to video understanding which combines shared temporal modelling of human actions with minimal overhead, to support multiple downstream tasks and enable cooperation when learning novel skills. We then propose EgoPack, a solution that creates a collection of task perspectives that can be carried across downstream tasks and used as a potential source of additional insights, as a backpack of skills that a robot can carry around and use when needed. We demonstrate the effectiveness and efficiency of our approach on four Ego4D benchmarks, outperforming current state-of-the-art methods., Comment: Accepted at IEEE/CVF Conference on Computer Vision and Pattern Recognition 2024. Project webpage at https://sapeirone.github.io/EgoPack

Recent transformer-based architectures have shown impressive results in the field of image segmentation. Thanks to their flexibility, they obtain outstanding performance in multiple segmentation tasks, such as semantic and panoptic, under a single unified framework. To achieve such impressive performance, these architectures employ intensive operations and require substantial computational resources, which are often not available, especially on edge devices. To fill this gap, we propose Prototype-based Efficient MaskFormer (PEM), an efficient transformer-based architecture that can operate in multiple segmentation tasks. PEM proposes a novel prototype-based cross-attention which leverages the redundancy of visual features to restrict the computation and improve the efficiency without harming the performance. In addition, PEM introduces an efficient multi-scale feature pyramid network, capable of extracting features that have high semantic content in an efficient way, thanks to the combination of deformable convolutions and context-based self-modulation. We benchmark the proposed PEM architecture on two tasks, semantic and panoptic segmentation, evaluated on two different datasets, Cityscapes and ADE20K. PEM demonstrates outstanding performance on every task and dataset, outperforming task-specific architectures while being comparable and even better than computationally-expensive baselines., Comment: CVPR 2024. Project page: https://niccolocavagnero.github.io/PEM

Many practical applications require training of semantic segmentation models on unlabelled datasets and their execution on low-resource hardware. Distillation from a trained source model may represent a solution for the first but does not account for the different distribution of the training data. Unsupervised domain adaptation (UDA) techniques claim to solve the domain shift, but in most cases assume the availability of the source data or an accessible white-box source model, which in practical applications are often unavailable for commercial and/or safety reasons. In this paper, we investigate a more challenging setting in which a lightweight model has to be trained on a target unlabelled dataset for semantic segmentation, under the assumption that we have access only to black-box source model predictions. Our method, named CoRTe, consists of (i) a pseudo-labelling function that extracts reliable knowledge from the black-box source model using its relative confidence, (ii) a pseudo label refinement method to retain and enhance the novel information learned by the student model on the target data, and (iii) a consistent training of the model using the extracted pseudo labels. We benchmark CoRTe on two synthetic-to-real settings, demonstrating remarkable results when using black-box models to transfer knowledge on lightweight models for a target data distribution., Comment: 11 pages, 6 figures, ICCV2023 workshop

Humans show an innate capability to identify tools to support specific actions. The association between objects parts and the actions they facilitate is usually named affordance. Being able to segment objects parts depending on the tasks they afford is crucial to enable intelligent robots to use objects of daily living. Traditional supervised learning methods for affordance segmentation require costly pixel-level annotations, while weakly supervised approaches, though less demanding, still rely on object-interaction examples and support a closed set of actions. These limitations hinder scalability, may introduce biases, and usually restrict models to a limited set of predefined actions. This paper proposes AffordanceCLIP, to overcome these limitations by leveraging the implicit affordance knowledge embedded within large pre-trained Vision-Language models like CLIP. We experimentally demonstrate that CLIP, although not explicitly trained for affordances detection, retains valuable information for the task. Our AffordanceCLIP achieves competitive zero-shot performance compared to methods with specialized training, while offering several advantages: i) it works with any action prompt, not just a predefined set; ii) it requires training only a small number of additional parameters compared to existing solutions and iii) eliminates the need for direct supervision on action-object pairs, opening new perspectives for functionality-based reasoning of models., Comment: Accepted to MAR Workshop at CVPR2024

Recent advancements in image segmentation have focused on enhancing the efficiency of the models to meet the demands of real-time applications, especially on edge devices. However, existing research has primarily concentrated on single-task settings, especially on semantic segmentation, leading to redundant efforts and specialized architectures for different tasks. To address this limitation, we propose a novel architecture for efficient multi-task image segmentation, capable of handling various segmentation tasks without sacrificing efficiency or accuracy. We introduce BiSeNetFormer, that leverages the efficiency of two-stream semantic segmentation architectures and it extends them into a mask classification framework. Our approach maintains the efficient spatial and context paths to capture detailed and semantic information, respectively, while leveraging an efficient transformed-based segmentation head that computes the binary masks and class probabilities. By seamlessly supporting multiple tasks, namely semantic and panoptic segmentation, BiSeNetFormer offers a versatile solution for multi-task segmentation. We evaluate our approach on popular datasets, Cityscapes and ADE20K, demonstrating impressive inference speeds while maintaining competitive accuracy compared to state-of-the-art architectures. Our results indicate that BiSeNetFormer represents a significant advancement towards fast, efficient, and multi-task segmentation networks, bridging the gap between model efficiency and task adaptability., Comment: Accepted to ECV workshop at CVPR2024

Soft robots are gaining popularity thanks to their intrinsic safety to contacts and adaptability. However, the potentially infinite number of Degrees of Freedom makes their modeling a daunting task, and in many cases only an approximated description is available. This challenge makes reinforcement learning (RL) based approaches inefficient when deployed on a realistic scenario, due to the large domain gap between models and the real platform. In this work, we demonstrate, for the first time, how Domain Randomization (DR) can solve this problem by enhancing RL policies for soft robots with: i) robustness w.r.t. unknown dynamics parameters; ii) reduced training times by exploiting drastically simpler dynamic models for learning; iii) better environment exploration, which can lead to exploitation of environmental constraints for optimal performance. Moreover, we introduce a novel algorithmic extension to previous adaptive domain randomization methods for the automatic inference of dynamics parameters for deformable objects. We provide an extensive evaluation in simulation on four different tasks and two soft robot designs, opening interesting perspectives for future research on Reinforcement Learning for closed-loop soft robot control., Comment: Presented as conference paper at IEEE/RSJ IROS 2023, Detroit, USA. Project website at https://andreaprotopapa.github.io/dr-soro

Deep neural networks for graphs have emerged as a powerful tool for learning on complex non-euclidean data, which is becoming increasingly common for a variety of different applications. Yet, although their potential has been widely recognised in the machine learning community, graph learning is largely unexplored for downstream tasks such as robotics applications. To fully unlock their potential, hence, we propose a review of graph neural architectures from a robotics perspective. The paper covers the fundamentals of graph-based models, including their architecture, training procedures, and applications. It also discusses recent advancements and challenges that arise in applied settings, related for example to the integration of perception, decision-making, and control. Finally, the paper provides an extensive review of various robotic applications that benefit from learning on graph structures, such as bodies and contacts modelling, robotic manipulation, action recognition, fleet motion planning, and many more. This survey aims to provide readers with a thorough understanding of the capabilities and limitations of graph neural architectures in robotics, and to highlight potential avenues for future research.

Neural Networks design is a complex and often daunting task, particularly for resource-constrained scenarios typical of mobile-sized models. Neural Architecture Search is a promising approach to automate this process, but existing competitive methods require large training time and computational resources to generate accurate models. To overcome these limits, this paper contributes with: i) a novel training-free metric, named Entropic Score, to estimate model expressivity through the aggregated element-wise entropy of its activations; ii) a cyclic search algorithm to separately yet synergistically search model size and topology. Entropic Score shows remarkable ability in searching for the topology of the network, and a proper combination with LogSynflow, to search for model size, yields superior capability to completely design high-performance Hybrid Transformers for edge applications in less than 1 GPU hour, resulting in the fastest and most accurate NAS method for ImageNet classification., Comment: 10 pages, 3 figures

To enable a safe and effective human-robot cooperation, it is crucial to develop models for the identification of human activities. Egocentric vision seems to be a viable solution to solve this problem, and therefore many works provide deep learning solutions to infer human actions from first person videos. However, although very promising, most of these do not consider the major challenges that comes with a realistic deployment, such as the portability of the model, the need for real-time inference, and the robustness with respect to the novel domains (i.e., new spaces, users, tasks). With this paper, we set the boundaries that egocentric vision models should consider for realistic applications, defining a novel setting of egocentric action recognition in the wild, which encourages researchers to develop novel, applications-aware solutions. We also present a new model-agnostic technique that enables the rapid repurposing of existing architectures in this new context, demonstrating the feasibility to deploy a model on a tiny device (Jetson Nano) and to perform the task directly on the edge with very low energy consumption (2.4W on average at 50 fps). The code is publicly available at: https://github.com/EgocentricVision/EgoWild., Comment: Accepted to RA-L, for associated video, see https://www.youtube.com/watch?v=7rtynmoYnuw&t=9s

In this report, we describe the technical details of our submission to the EPIC-Kitchens-100 Unsupervised Domain Adaptation (UDA) Challenge in Action Recognition. To tackle the domain-shift which exists under the UDA setting, we first exploited a recent Domain Generalization (DG) technique, called Relative Norm Alignment (RNA). Secondly, we extended this approach to work on unlabelled target data, enabling a simpler adaptation of the model to the target distribution in an unsupervised fashion. To this purpose, we included in our framework UDA algorithms, such as multi-level adversarial alignment and attentive entropy. By analyzing the challenge setting, we notice the presence of a secondary concurrence shift in the data, which is usually called environmental bias. It is caused by the existence of different environments, i.e., kitchens. To deal with these two shifts (environmental and temporal), we extended our system to perform Multi-Source Multi-Target Domain Adaptation. Finally, we employed distinct models in our final proposal to leverage the potential of popular video architectures, and we introduced two more losses for the ensemble adaptation. Our submission (entry 'plnet') is visible on the leaderboard and ranked in 2nd position for 'verb', and in 3rd position for both 'noun' and 'action'., Comment: 3rd place in the 2022 EPIC-KITCHENS-100 Unsupervised Domain Adaptation Challenge for Action Recognition. arXiv admin note: substantial text overlap with arXiv:2107.00337

In the last decade, most research in Machine Learning contributed to the improvement of existing models, with the aim of increasing the performance of neural networks for the solution of a variety of different tasks. However, such advancements often come at the cost of an increase of model memory and computational requirements. This represents a significant limitation for the deployability of research output in realistic settings, where the cost, the energy consumption, and the complexity of the framework play a crucial role. To solve this issue, the designer should search for models that maximise the performance while limiting its footprint. Typical approaches to reach this goal rely either on manual procedures, which cannot guarantee the optimality of the final design, or upon Neural Architecture Search algorithms to automatise the process, at the expenses of extremely high computational time. This paper provides a solution for the fast identification of a neural network that maximises the model accuracy while preserving size and computational constraints typical of tiny devices. Our approach, named FreeREA, is a custom cell-based evolution NAS algorithm that exploits an optimised combination of training-free metrics to rank architectures during the search, thus without need of model training. Our experiments, carried out on the common benchmarks NAS-Bench-101 and NATS-Bench, demonstrate that i) FreeREA is a fast, efficient, and effective search method for models automatic design; ii) it outperforms State of the Art training-based and training-free techniques in all the datasets and benchmarks considered, and iii) it can easily generalise to constrained scenarios, representing a competitive solution for fast Neural Architecture Search in generic constrained applications. The code is available at \url{https://github.com/NiccoloCavagnero/FreeREA}., Comment: 16 pages, 4 figures

Physics simulators have shown great promise for conveniently learning reinforcement learning policies in safe, unconstrained environments. However, transferring the acquired knowledge to the real world can be challenging due to the reality gap. To this end, several methods have been recently proposed to automatically tune simulator parameters with posterior distributions given real data, for use with domain randomization at training time. These approaches have been shown to work for various robotic tasks under different settings and assumptions. Nevertheless, existing literature lacks a thorough comparison of existing adaptive domain randomization methods with respect to transfer performance and real-data efficiency. In this work, we present an open benchmark for both offline and online methods (SimOpt, BayRn, DROID, DROPO), to shed light on which are most suitable for each setting and task at hand. We found that online methods are limited by the quality of the currently learned policy for the next iteration, while offline methods may sometimes fail when replaying trajectories in simulation with open-loop commands. The code used will be released at https://github.com/gabrieletiboni/adr-benchmark., Comment: 15 pages, 6 figures

© 2020 IEEE. This paper tackles the challenge of predicting grasp failures in soft hands before they happen, by combining deep learning with a sensing strategy based on distributed Inertial Measurement Units. We propose two neural architectures, which we implemented and tested with an articulated soft hand-the Pisa/IIT SoftHand-and a continuously deformable soft hand-the RBO Hand. The first architecture (Classifier) implements a-posteriori detection of the failure event, serving as a test-bench to assess the possibility of extracting failure information from the discussed input signals. This network reaches up to 100% of accuracy within our experimental validation. Motivated by these results, we introduce a second architecture (Predictor), which is the main contribution of the paper. This network works on-line and takes as input a multidimensional continuum stream of raw signals coming from the Inertial Measurement Units. The network is trained to predict the occurrence in the near future of a failure event. The Predictor detects 100% of failures with both hands, with the detection happening on average 1.96 seconds before the actual failing occurs-leaving plenty of time to an hypothetical controller to react.

© 2020 IEEE. Physical interaction of robots with their environment is a challenging problem because of the exchanged forces. Hybrid position/force control schemes often exhibit problems during the contact phase, whereas impedance control appears to be more simple and reliable, especially when impedance is shaped to be energetically passive. Even if recent technologies enable shaping the impedance of a robot, how best to plan impedance parameters for task execution remains an open question. In this paper we present an optimization-based approach to plan not only the robot motion but also its desired end-effector mechanical impedance. We show how our methodology is able to take into account the transition from free motion to a contact condition, typical of physical interaction tasks. Results are presented for planar and three-dimensional open-chain manipulator arms. The compositionality of mechanical impedance is exploited to deal with kinematic redundancy and multi-arm manipulation., U.S. National Science Foundation (Grants 1637824, 1724135, and 1826097)

Deep Neural Networks (DNNs) enable a wide series of technological advancements, ranging from clinical imaging, to predictive industrial maintenance and autonomous driving. However, recent findings indicate that transient hardware faults may corrupt the models prediction dramatically. For instance, the radiation-induced misprediction probability can be so high to impede a safe deployment of DNNs models at scale, urging the need for efficient and effective hardening solutions. In this work, we propose to tackle the reliability issue both at training and model design time. First, we show that vanilla models are highly affected by transient faults, that can induce a performances drop up to 37%. Hence, we provide three zero-overhead solutions, based on DNN re-design and re-train, that can improve DNNs reliability to transient faults up to one order of magnitude. We complement our work with extensive ablation studies to quantify the gain in performances of each hardening component., Comment: 7 pages, 6 figures

The need for users’ safety and technology acceptability has incredibly increased with the deployment of co-bots physically interacting with humans in industrial settings, and for people assistance. A well-studied approach to meet these requirements is to ensure human-like robot motions and interactions. In this manuscript, we present a research approach that moves from the understanding of human movements and derives usefull guidelines for the planning of arm movements and the learning of skills for physical interaction of robots with the surrounding environment., Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Learning & Autonomous Control

Learning & Autonomous Control

In recent years, the spread of data-driven approaches for robotic grasp synthesis has come with the increasing need for reliable datasets, which can be built e.g. through video labelling. To this goal, it is important to define suitable rules to characterize the main human grasp types, for easily identifying them in video streams. In this work, we present a novel taxonomy that builds upon the related state of the art, but it is specifically thought for video labelling. It focuses on the interaction of the hand with the environment and accounts for pre-contact phases, bi-manual grasps as well as non-prehensile strategies. This study is complemented with a dataset of labelled videos of subjects performing activities of daily living, for a total of nine hours, and the description of MatLab tools for labelling new videos. Both hands were labelled at any time. We used these labelled data for performing a preliminary statistical description of the occurrences of the here proposed class types., Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Learning & Autonomous Control

The need for users’ safety and technology acceptability has incredibly increased with the deployment of co-bots physically interacting with humans in industrial settings, and for people assistance. A well-studied approach to meet these requirements is to ensure human-like robot motions and interactions. In this manuscript, we present a research approach that moves from the understanding of human movements and derives usefull guidelines for the planning of arm movements and the learning of skills for physical interaction of robots with the surrounding environment., Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Learning & Autonomous Control

Learning & Autonomous Control

In recent years, the spread of data-driven approaches for robotic grasp synthesis has come with the increasing need for reliable datasets, which can be built e.g. through video labelling. To this goal, it is important to define suitable rules to characterize the main human grasp types, for easily identifying them in video streams. In this work, we present a novel taxonomy that builds upon the related state of the art, but it is specifically thought for video labelling. It focuses on the interaction of the hand with the environment and accounts for pre-contact phases, bi-manual grasps as well as non-prehensile strategies. This study is complemented with a dataset of labelled videos of subjects performing activities of daily living, for a total of nine hours, and the description of MatLab tools for labelling new videos. Both hands were labelled at any time. We used these labelled data for performing a preliminary statistical description of the occurrences of the here proposed class types., Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public., Learning & Autonomous Control

BACKGROUND: Human-likeliness of robot movements is a key component to enable a safe and effective human-robot interaction, since it contributes to increase acceptance and motion predictability of robots that have to closely interact with people, e.g. for assistance and rehabilitation purposes. Several parameters have been used to quantify how much a robot behaves like a human, which encompass aspects related to both the robot appearance and motion. The latter point is fundamental to allow the operator to interpret robotic actions, and plan a meaningful reactions. While different approaches have been presented in literature, which aim at devising bio-aware control guidelines, a direct implementation of human actions for robot planning is not straightforward, still representing an open issue in robotics. METHODS: We propose to embed a synergistic representation of human movements for robot motion generation. To do this, we recorded human upper-limb motions during daily living activities. We used functional Principal Component Analysis (fPCA) to extract principal motion patterns. We then formulated the planning problem by optimizing the weights of a reduced set of these components. For free-motions, our planning method results into a closed form solution which uses only one principal component. In case of obstacles, a numerical routine is proposed, incrementally enrolling principal components until the problem is solved with a suitable precision. RESULTS: Results of fPCA show that more than 80% of the observed variance can be explained by only three functional components. The application of our method to different meaningful movements, with and without obstacles, show that our approach is able to generate complex motions with a very reduced number of functional components. We show that the first synergy alone accounts for the 96% of cost reduction and that three components are able to achieve a satisfactory motion reconstruction in all the considered cases. CONCLUSIONS