Teaching heterogeneous robotic cells is difficult and time-consuming. We present a context-aware robot teaching solution based on an open-source framework that decouples the task instruction from the task execution to simplify the teaching workflow significantly. To demonstrate the advantages of the solution, a human operator uses a VR headset and controllers to teach a virtual robot to perform a task, e.g., pick-and-place in a virtual world. The task is automatically adapted to different settings of operation cells and executed by physical robots located at different geographical locations.

Feature attribution a.k.a.input salience methods which assign an importance score to a feature are abundant but may produce surprisingly different results for the same model on the same input.While differences are expected if disparate definitions of importance are assumed, most methods claim to provide faithful attributions and point at the features most relevant for a model's prediction.Existing work on faithfulness evaluation is not conclusive and does not provide a clear answer as to how different methods are to be compared.Focusing on text classification and the model debugging scenario, our main contribution is a protocol for faithfulness evaluation that makes use of partially synthetic data to obtain ground truth for feature importance ranking.Following the protocol, we do an in-depth analysis of four standard salience method classes on a range of datasets and lexical shortcuts for BERT and LSTM models.We demonstrate that some of the most popular method configurations provide poor results even for simple shortcuts while a method judged to be too simplistic works remarkably well for BERT.

Even though model-driven engineering reduces complexity during the development of self-adaptive systems and [email protected] enables using them during runtime, connecting models to different external systems still involves manual work. Those connections are essential to the complete system, as they enable external systems to react to changes in the internal model and vice versa. In our case, the model is based on Relational Reference Attribute Grammars, an extension of Attribute Grammars to enable conceptual models at runtime while retaining their benefits of modular specification and an incremental evaluation scheme. We present an approach to enable concise specification of the causal connection and needed transformations to match required formats or semantics. To show its applicability, a case study showing the coordination of multiple industrial robot arms using models is presented. We show that using our approach, connections can be specified more concisely while maintaining the same efficiency as hand-written code. The artefact comprising all source code and an executable version of the case studies is available at https://doi.org/10.5281/zenodo.7009758.

Collaborative robotic cells combine human skills with the latest advancements in robotic accuracy and reliability. Cobotic cell parts are distributed and adapt their behavior to changing tasks and environments. The specific missions of cobotic cells, depend on their field of application, but are critical for human safety, which introduces complexity, increasing testing and development effort. Component-based software engineering is used to manage complexity, but ensuring safety and correctness requires verification and validation, which is complex and demanding to re-ensure, when composed behavior changes. This also applies to the widely used middleware Robot Operating System (ROS), where existing approaches only model high level communication or integrate models. Also, verification of cobotic cells must reflect their context-adaptivity, to check safety critical reactions to contexts-changes. To overcome these inhibitors, a model-driven development approach based on Petri nets is proposed, modeling central aspects of ROS-based cobotic cells. By using formal models, the testing effort at development time is reduced, because global behavior remains formally proven, and only local components have to be retested. Within this work, the plans for this model-driven software approach are reported.

Gorman and Bedrick (2019) argued for using random splits rather than standard splits in NLP experiments.We argue that random splits, like standard splits, lead to overly optimistic performance estimates.We can also split data in biased or adversarial ways, e.g., training on short sentences and evaluating on long ones.Biased sampling has been used in domain adaptation to simulate real-world drift; this is known as the covariate shift assumption.In NLP, however, even worst-case splits, maximizing bias, often under-estimate the error observed on new samples of in-domain data, i.e., the data that models should minimally generalize to at test time.This invalidates the covariate shift assumption.Instead of using multiple random splits, future benchmarks should ideally include multiple, independent test sets instead; if infeasible, we argue that multiple biased splits leads to more realistic performance estimates than multiple random splits.

Recently more attention has been given to adversarial attacks on neural networks for natural language processing (NLP).A central research topic has been the investigation of search algorithms and search constraints, accompanied by benchmark algorithms and tasks.We implement an algorithm inspired by zeroth order optimization-based attacks and compare with the benchmark results in the TextAttack framework.Surprisingly, we find that optimizationbased methods do not yield any improvement in a constrained setup and slightly benefit from approximate gradient information only in unconstrained setups where search spaces are larger.In contrast, simple heuristics exploiting nearest neighbors without querying the target function yield substantial success rates in constrained setups, and nearly full success rate in unconstrained setups, at an order of magnitude fewer queries.We conclude from these results that current TextAttack benchmark tasks are too easy and constraints are too strict, preventing meaningful research on black-box adversarial text attacks.

Embeddings are generic representations that are useful for many NLP tasks. In this paper, we introduce DENSIFIER, a method that learns an orthogonal transformation of the embedding space that focuses the information relevant for a task in an ultradense subspace of a dimensionality that is smaller by a factor of 100 than the original space. We show that ultradense embeddings generated by DENSIFIER reach state of the art on a lexicon creation task in which words are annotated with three types of lexical information - sentiment, concreteness and frequency. On the SemEval2015 10B sentiment analysis task we show that no information is lost when the ultradense subspace is used, but training is an order of magnitude more efficient due to the compactness of the ultradense space.

We put forward the hypothesis that highaccuracy sentiment analysis is only possible if word senses with different polarity are accurately recognized.We provide evidence for this hypothesis in a case study for the adjective "hard" and propose contextually enhanced sentiment lexicons that contain the information necessary for sentiment-relevant sense disambiguation.An experimental evaluation demonstrates that senses with different polarity can be distinguished well using a combination of standard and novel features.

In this paper we introduce a new layer for the task of handwriting recognition. We add semantic information by means of ontologies. The task of our recognizer therefore is not only to recognize the ASCII transcription of the handwritten document, but also to identify the semantic concepts which appear in the text. This task is called ontology-based information extraction (OBIE), which has been applied to electronic documents recently. OBIE methods first segment the text into tokens, then identify their values and their corresponding instances of the ontology, and finally try to generate new facts based on the text. To the authors' knowledge, in this paper OBIE is proposed for the first time in handwriting literature. In our experiments we have evaluated the process up to the instantiation. We have found that using not only the top alternative, but also the k-best alternatives increases the performance of information extraction. Furthermore, the use of an ontology-based lexicon results in another performance increase.