Search

MU-MIMO is a high-speed technique in IEEE 802.11ac and upcoming ax technologies that improves spectral efficiency by allowing concurrent communication between one Access Point and multiple users. In this paper, we present LATTE, a novel framework that proposes MU-MIMO-aware optimization for multi-user multimedia applications over IEEE 802.11ac/ax. Taking a cross-layer approach, LATTE first optimizes the MU-MIMO user group selection for the users with the same characteristics in the PHY/MAC layer. It then optimizes the video bitrate for each group accordingly. We present our design and its evaluation on smartphones and laptops over 802.11ac WiFi. Our experimental evaluations indicate that LATTE can outperform other video rate adaptation algorithms.

Software Defined Networking (SDN) has emerged as a new technology that enabled dynamic adaptation and configuration to the traditional networking. In this work, using SDN centralized control, we propose a path algorithm to avoid congestion in data communication networks. Our proposed algorithm will find a path between the source and the destination nodes by considering hop counts and the available bandwidth among the links. The path we find can be slightly longer, but is less congested compared to the congested least hops path. The found path will lead to better traffic load distribution in the network. We implemented the proposed algorithm and carried out experiments in Mininet using RYU controller. Our algorithm shows a good performance improvement in resolving the congestion when compared to the existing path algorithm in the RYU controller.

The evolution of malware has long plagued machine learning-based detection systems, as malware authors develop innovative strategies to evade detection and chase profits. This induces concept drift as the test distribution diverges from the training, causing performance decay that requires constant monitoring and adaptation. In this work, we analyze the adaptation strategy used by DroidEvolver, a state-of-the-art learning system that self-updates using pseudo-labels to avoid the high overhead associated with obtaining a new ground truth. After removing sources of experimental bias present in the original evaluation, we identify a number of flaws in the generation and integration of these pseudo-labels, leading to a rapid onset of performance degradation as the model poisons itself. We propose DroidEvolver++, a more robust variant of DroidEvolver, to address these issues and highlight the role of pseudo-labels in addressing concept drift. We test the tolerance of the adaptation strategy versus different degrees of pseudo-label noise and propose the adoption of methods to ensure only high-quality pseudo-labels are used for updates. Ultimately, we conclude that the use of pseudo-labeling remains a promising solution to limitations on labeling capacity, but great care must be taken when designing update mechanisms to avoid negative feedback loops and self-poisoning which have catastrophic effects on performance.

Edge video analytics based on deep learning has become an important building block for many modern intelligent applications such as mobile augmented reality and autonomous driving. Various mechanisms have been developed to handle dynamic wireless networks, compute resource availability, and achieve high analytics accuracy via filtering, DNN compression, pruning, and adaptation. So far, limited attention has been paid to timeliness - -providing strict service-level objectives (SLO) for edge video analytics pipelines, which is essential for the usability of user-interactive and mission-critical intelligent applications. In this paper, we analyze the challenges in achieving SLO for edge video analytics and present a system design for timely edge video analytics over the air leveraging a simple yet effective idea - -feedback control. Our preliminary evaluation based on a system prototype and real-world network traces shows the potential of our design. We also discuss the limitations, calling for future work.

Sign languages (LS) are the natural form of communication between deaf people. However, although sign languages are recognized in many countries as a language, access to information and education is still a common problem for deaf people. In Brazil, despite the existence of several accessibility laws that determine support for deaf people, many are still dependent on Libras interpreters to communicate, which impairs their autonomy. As a result, there is a great difficulty for the deaf to exercise their citizenship and have access to quality information and education through sign languages, which generally implies linguistic delay and knowledge acquisition. In this sense, the present work addresses a proposal for an accessible interface of tele-education for the hearing impaired in order to facilitate access to information, making it possible to bring the figure of the remote interpreter to the classroom. This solution involves adapting the interface of the open source system, making it possible for the remote interpreter to be present to meet the needs of deaf students and make the classroom experience more accessible. In addition, a solution for displaying live subtitles was also implemented. The results obtained in this study proved to be promising, with 50% of users successfully completing more than 93.33% of the requested tasks. In addition, the group with the proposed solution was less dispersed when compared to the other group, concentrating most of its results above 80%, in a way that leads to believe that this type of adaptation can be welcomed by deaf users in the context of tele-education.

The presentation of sensory effects in sync with 360° content has the potential to increase user immersion. However, the authoring process for such effects is laborious and slow. It requires the author to specify in time and space the presentation characteristics of effects. This paper contributes to address this need with a generic architecture for inferring the presentation characteristics of sensory effects in 360° content. The proposed architecture decouples and eases the adaptation of multimedia content analysis techniques for recognizing sensory effects. We highlight the original contribution of this work of presenting the automatic recognition of sensory effects location in 360° content.

Although existing work has demonstrated the feasibility of streaming volumetric content to a single user, there exist many appealing applications (e.g., classroom education and collaborative design) that involve multiple users who watch the same volumetric content simultaneously. In this paper, we first perform a scaling experiment to demonstrate the challenges of streaming high-quality volumetric videos to multiple users and reveal the viewport-similarity opportunity that we can leverage to effectively optimize the network resource utilization using multicast over mmWave. We then develop a holistic research agenda for improving the performance and quality of experience for multi-user volumetric video streaming on commodity devices. Our proposed research includes joint viewport prediction and blockage mitigation for multiple users, multicast grouping based on viewport similarity, customized mmWave beam design for efficient multicast, and mmWave-aware multi-user video rate adaptation. Finally, we discuss the open challenges of building a practical system with the proposed research roadmap.

Click-through rate (CTR) prediction is crucial for online sponsored search advertising. Several successful CTR models have been adopted in the industry, including the regularized logistic regression (LR). Nonetheless, the learning process suffers from two limitations: 1) Feature crosses for high-order information may generate trillions of features, which are sparse for online learning examples; 2) Rapid changing of data distribution brings challenges to the accurate learning since the model has to perform a fast adaptation on the new data. Moreover, existing adaptive optimizers are ineffective in handling the sparsity issue for high-dimensional features. In this paper, we propose to learn an optimizer in a meta-learning scenario, where the optimizer is learned on prior data and can be easily adapted to the new data. We firstly build a low-dimensional feature embedding on prior data to encode the association among features. Then, the gradients on new data can be decomposed into the low-dimensional space, enabling the parameter update smoothed and relieving the sparsity. Note that this technology could be deployed into a distributed system to ensure efficient online learning on the trillions-level parameters. We conduct extensive experiments to evaluate the algorithm in terms of prediction accuracy and actual revenue. Experimental results demonstrate that the proposed framework achieves a promising prediction on the new data. The final online revenue is noticeably improved compared to the baseline. This framework was initially deployed in Baidu Search Ads (a.k.a. Phoenix Nest) in 2014 and is currently still being used in certain modules of Baidu's ads systems.

Graph Neural Networks (GNNs) have emerged as a series of effective learning methods for graph-related tasks. However, GNNs are shown vulnerable to adversarial attacks, where attackers can fool GNNs into making wrong predictions on adversarial samples with well-designed perturbations. Specifically, we observe that the current evasion attacks suffer from two limitations: (1) the attack strategy based on the reinforcement learning method might not be transferable when the attack budget changes; (2) the greedy mechanism in the vanilla gradient-based method ignores the long-term benefits of each perturbation operation. In this paper, we propose a new attack method named projective ranking to overcome the above limitations. Our idea is to learn a powerful attack strategy considering the long-term benefits of perturbations, then adjust it as little as possible to generate adversarial samples under different budgets. We further employ mutual information to measure the long-term benefits of each perturbation and rank them accordingly, so the learned attack strategy has better attack performance. Our method dramatically reduces the adaptation cost of learning a new attack strategy by projecting the attack strategy when the attack budget changes. Our preliminary evaluation results in synthesized and real-world datasets demonstrate that our method owns powerful attack performance and effective transferability.

Practical recommender systems experience a cold-start problem when observed user-item interactions in the history are insufficient. Meta learning, especially gradient based one, can be adopted to tackle this problem by learning initial parameters of the model and thus allowing fast adaptation to a specific task from limited data examples. Though with significant performance improvement, it commonly suffers from two critical issues: the non-compatibility with mainstream industrial deployment and the heavy computational burdens, both due to the inner-loop gradient operation. These two issues make them hard to be applied in practical recommender systems. To enjoy the benefits of meta learning framework and mitigate these problems, we propose a recommendation framework called Contextual Modulation Meta Learning (CMML). CMML is composed of fully feed-forward operations so it is computationally efficient and completely compatible with the mainstream industrial deployment. CMML consists of three components, including a context encoder that can generate context embedding to represent a specific task, a hybrid context generator that aggregates specific user-item features with task-level context, and a contextual modulation network, which can modulate the recommendation model to adapt effectively. We validate our approach on both scenario-specific and user-specific cold-start setting on various real-world datasets, showing CMML can achieve comparable or even better performance with gradient based methods yet with much higher computational efficiency and better interpretability., corresponding to ; Accepted by CIKM 2021

The wide adaptation of the Semantic Web and the Resource Description Framework (RDF) has made available many important datasets. The SPARQL query language facilitates the exploration of this information, which is available in a semi-structured way that does not comply with relational data models, deviating from exploration techniques that most researchers are familiar with. Usually, only people with training and extensive knowledge of the RDF model can explore and understand it in depth. We present here a platform that supports the users with querying, exploring and visualizing information available in SPARQL endpoints. A dedicated visualization module, built upon a knowledge database, allows us to provide case-specific visualization solutions for SPARQL query results. The selection is based exclusively on features extracted from the result, without any knowledge about the structure, content and characteristics of the underlying dataset.

Human interactions with items are being constantly logged, which enables advanced representation learning and facilitates various tasks. Instead of generating static embeddings at the end of training, several temporal embedding methods were recently proposed to learn user and item embeddings as functions of time, where each entity has a trajectory of embedding vectors aiming to encode the full dynamics. However, these methods may not be optimal to encode the dynamical behaviors on the interaction graphs in that they can not generate "fully''-temporal embeddings and do not consider information propagation. In this paper, we tackle the issues and propose CoPE (Co ntinuous P ropagation and E volution). We use an ordinary differential equation based graph neural network to model information propagation and more sophisticated evolution patterns. We train CoPE on sequences of interactions with the help of meta-learning to ensure fast adaptation to the most recent interactions. We evaluate CoPE on three tasks and prove its effectiveness.

Adapting neural networks to unseen tasks with few training samples on resource-constrained devices benefits various Internet-of-Things applications. Such neural networks should learn the new tasks in few shots and be compact in size. Meta-learning enables few-shot learning, yet the meta-trained networks can be over-parameterised. However, naive combination of standard compression techniques like network pruning with meta-learning jeopardises the ability for fast adaptation. In this work, we propose adaptation-aware network pruning (ANP), a novel pruning scheme that works with existing meta-learning methods for a compact network capable of fast adaptation. ANP uses weight importance metric that is based on the sensitivity of the meta-objective rather than the conventional loss function, and adopts approximation of derivatives and layer-wise pruning techniques to reduce the overhead of computing the new importance metric. Evaluations on few-shot classification benchmarks show that ANP can prune meta-trained convolutional and residual networks by 85% without affecting their fast adaptation.

In many high-stakes applications of machine learning models, outputting only predictions or providing statistical confidence is usually insufficient to gain trust from end users, who often prefer a transparent reasoning paradigm. Despite the recent encouraging developments on deep networks for sequential data modeling, due to the highly recursive functions, the underlying rationales of their predictions are difficult to explain. Thus, in this paper, we aim to develop a sequence modeling approach that explains its own predictions by breaking input sequences down into evidencing segments (i.e., sub-sequences) in its reasoning. To this end, we build our model upon convolutional neural networks, which, in their vanilla forms, associates local receptive fields with outputs in an obscure manner. To unveil it, we resort to case-based reasoning, and design prototype modules whose units (i.e., prototypes) resemble exemplar segments in the problem domain. Each prediction is obtained by combining the comparisons between the prototypes and the segments of an input. To enhance interpretability, we propose a training objective that delicately adapts the distribution of prototypes to the data distribution in latent spaces, and design an algorithm to map prototypes to human-understandable segments. Through extensive experiments in a variety of domains, we demonstrate that our model can achieve high interpretability generally, together with a competitive accuracy to the state-of-the-art approaches.

We address the problem of low resource machine learning in the form of few-shot learning (FSL) applied to word recognition in both mono-lingual and cross-lingual settings. Recently, we proposed an adaptation of a FSL framework, matching networks (MN) to a suite of speech recognition tasks such as multi-speaker small-to-medium vocabulary word recognition and frame-wise phoneme recognition tasks under mel-spectrogram and single-frame feature representations. In this paper, we extend this FSL adaptation of MN to multi-speaker isolated word recognition (IWR), in a framework termed MN-IWR. The IWR task is specifically set in a ‘command-and-control’ (C&C) scenario with the requirement of needing only very few-shot examples (e.g. up to 20) for a target IWR classification task with vocabularies defined dynamically. Moreover, our proposed MN-IWR framework addresses a cross-domain and cross-lingual setting defined as below: a model is trained on a possibly large set of words in a source-language and used for inference on a cross-domain task (vocabulary of words different from the training vocabulary) or a cross-lingual task (vocabulary of words from a target-language different from the source-language). In this work, we present the main formulation of the MN-IWR framework, its adaptation from source-to-target tasks and results on TIMIT vocabulary of words in a mono-lingual setting and on English, Kannada and Tamil words in cross-lingual settings and report very high performances of the proposed MN-IWR FSL paradigm over conventional IWR classification without the FSL advantage of the MN formulation.

The development and use of intelligent agents are currently growing vertiginously. This growth has raised awareness of big technology companies that launched platforms that ease chatbot development. However, chatbots still fail to meet users' social expectations associated with cultural values imbued during the chatbot development (concerned or not with anthropomorphic topics). In literature, nowadays, there is a lack of research on culture in chatbot building platforms, primarily focusing on non-programmers. This research investigates interactive resources for cultural adaptation on the Blip Builder platform based on Semiotic Engineering. We evaluate the Blip communicability using the Semiotic Inspection Method combined with the Cultural Viewpoint Metaphors, followed by an autoethnography step in the design process of a conversational agent based on Omotenashi, a Japanese cultural characteristic. The main results indicate that the platform offers opportunities that enable cultural adaptation, mainly improving the adaptation of the chatbot behavior to the Omotenashi properties. But also convey some barriers for cultural expression.

The development of expressive embodied conversational agent (ECA) still remains a big challenge. During an interaction partners continuously adapt their behaviors one to the other [7]. Adaptation mechanisms may take different forms such as the choice of same vocabulary and grammatical form [31], imitation and synchronization [7]. The aim of my PhD project is to improve the interaction between human and agent. The key idea is to create an interactive loop between human and agent which allows the virtual agent to continuously adapt its behavior according to its partner’s behavior. The main idea is to learn how dyad of humans adapt their behaviors and implement it into human-agent interaction. My work, based on recurrent neural network, focuses on nonverbal behavior generation and addresses several scientific locks like the multimodality, the intra-personal temporality of multimodal signals or the temporality between partner’s social cues. We plan to build a model learned in an end-to-end fashion that generates behaviors considering both acoustic and visual modalities.

Deep\footnote learning-based semantic segmentation methods require a huge amount of training images with pixel-level annotations. Unsupervised domain adaptation (UDA) for semantic segmentation enables transferring knowledge learned from the synthetic data (source domain) with low-cost annotations to the real images (target domain). However, current UDA methods mostly require full access to the source domain data for feasible adaptation, which limits their applications in real-world scenarios with privacy, storage, or transmission issues. To this end, this paper identifies and addresses a more practical but challenging problem of UDA for semantic segmentation, where access to the original source domain data is forbidden. In other words, only the pre-trained source model and unlabelled target domain data are available for adaptation. To tackle the problem, we propose to construct a set of source domain virtual data to mimic the source domain distribution by identifying the target domain high-confidence samples predicted by the pre-trained source model. Then by analyzing the data properties in the cross-domain semantic segmentation tasks, we propose an uncertainty and prior distribution-aware domain adaptation method to align the virtual source domain and the target domain with both adversarial learning and self-training strategies. Extensive experiments on three cross-domain semantic segmentation datasets with in-depth analyses verify the effectiveness of the proposed method.

Cross-domain Facial Expression Recognition (FER) is challenging due to the difficulty of concurrently handling the domain shift and semantic gap during domain adaptation. Existing methods mainly focus on reducing the domain discrepancy for transferable features but fail to decrease the semantic one, which may result in negative transfer. To this end, we propose Joint Discriminative and Mutual Adaptation Networks (JDMAN), which collaboratively bridge the domain shift and semantic gap by domain- and category-level co-adaptation based on mutual information and discriminative metric learning techniques. Specifically, we design a mutual information minimization module for domain-level adaptation, which narrows the domain shift by simultaneously distilling the domain-invariant components and eliminating the untransferable ones lying in different domains. Moreover, we propose a semantic metric learning module for category-level adaptation, which can close the semantic discrepancy during discriminative intra-domain representation learning and transferable inter-domain knowledge discovery. These two modules are jointly leveraged in our JDMAN to safely transfer the source knowledge to target data in an end-to-end manner. Extensive experimental results on six databases show that our method achieves state-of-the-art performance. The code of our JDMAN is available at https://github.com/YingjianLi/JDMAN.

MADMAX is a Haskell-embedded DSL for multi-attribute, multi-layered decision making. An important feature of this DSL is the ability to generate explanations of why a computed optimal solution is better than its alternatives. The functional approach and Haskell's type system support a high-level formulation of decision-making problems, which facilitates a number of innovations, including the gradual evolution and adaptation of problem representations, a more user-friendly form of sensitivity analysis based on problem domain data, and fine-grained control over explanations.

Tile-based approach is widely adopted in adaptive 360\textdegree~video streaming systems. Existing QoE-driven streaming approaches usually obtain the tile selection and adjust the bitrate based on the viewport prediction with a fixed tiling, which fail to consider the unstable prediction performance. However, varying the tiling of the video can produce different number of tiles with different sizes, and thus can have distinct impacts on error tolerance for viewport prediction and on decoding complexity for resource-constrained mobile client. In this work, we introduce adaptive tiling into the conventional bitrate adaptation for mobile 360degree~video streaming. We first analyze the impacts of tilings on tile selection and decoding time, which verify the benefit of tiling adaptation in various practical aspects. We then formulate the QoE optimization problem for adaptive tiling and bitrate streaming and discuss the design details of our adaptation algorithm, which can adapt to the performance of viewport prediction and the decoding capabilities of mobile clients in addition to the conventional influencing factors. Finally, the superiority of our proposed approach compared with the state-of-the-art methods is evaluated through extensive trace-driven simulations.

Video conferencing applications have seen explosive growth both in the number of available applications and their use. However, there have been few studies on the detailed analysis of video conferencing applications with respect to network dynamics, yet understanding these dynamics is essential for network design and improving these applications. In this paper, we carry out an in-depth measurement and modeling study on the rate control algorithms used in six popular commercial video conferencing applications. Based on macroscopic behaviors commonly observed across these applications in our extensive measurements, we construct a unified architecture to model the rate control mechanisms of individual applications. We then reconstruct each application's rate control by inferring key parameters that closely follow its rate control and quality adaptation behaviors. To our knowledge, this is the first work that reverse-engineers rate control algorithms of popular video conferencing applications, which are often unknown or hidden as they are proprietary software. We confirm our analysis and models using an end-to-end testbed that can capture the dynamics of each application under a variety of network conditions. We also show how we can use these models to gain insights into the particular behaviors of an application in two practical scenarios.

Recent advances in unsupervised domain adaptation have achieved remarkable performance on semantic segmentation tasks. Despite such progress, existing works mainly focus on bridging the inter-domain gaps between the source and target domain, while only few of them noticed the intra-domain gaps within the target data. In this work, we propose a pixel-level intra-domain adaptation approach to reduce the intra-domain gaps within the target data. Compared with image-level methods, ours treats each pixel as an instance, which adapts the segmentation model at a more fine-grained level. Specifically, we first conduct the inter-domain adaptation between the source and target domain; Then, we separate the pixels in target images into the easy and hard subdomains; Finally, we propose a pixel-level adversarial training strategy to adapt a segmentation network from the easy to the hard subdomain. Moreover, we show that the segmentation accuracy can be further improved by incorporating a continuous indexing technique in the adversarial training. Experimental results show the effectiveness of our method against existing state-of-the-art approaches.

Directly deploying a trained multi-modal classifier to a new environment usually leads to poor performance due to the well-known domain shift problem. Existing multi-modal domain adaptation methods treated each modality equally and optimize the sub-models of different modalities synchronously. However, as observed in this paper, the degrees of domain shift in different modalities are usually diverse. We propose a novel Differentiated Learning framework to make use of the diversity between multiple modalities for more effective domain adaptation. Specifically, we model the classifiers of different modalities as a group of teacher/student sub-models, and a novel Prototype based Reliability Measurement is presented to estimate the reliability of the recognition results made by each sub-model on the target domain. More reliable results are then picked up as teaching materials for all sub-models in the group. Considering the diversity of different modalities, each sub-model performs the Asynchronous Curriculum Learning by choosing the teaching materials from easy to hard measured by itself. Furthermore, a reliability-aware fusion scheme is proposed to combine all optimized sub-models to support final decision. Comprehensive experiments based on three multi-modal datasets with different learning tasks have been conducted, which show the superior performance of our model while comparing with state-of-the-art multi-modal domain adaptation models.

This paper proposes an adaptation of session types to provide behavioural information about public functions in Elixir modules. We formalise typechecking rules for the main constructs of the language. This allows us to statically determine whether a function implementation observes its session endpoint specification. Based on this type system, we then construct a tool that automates typechecking for Elixir modules.

Brightening low-light images of diverse scenes is a challenging but widely concerned task in the multimedia community. Convolutional Neural Networks (CNNs) based approaches mostly acquire the enhanced model by learning the data distribution from the specific scenes. However, these works present poor adaptability (even fail) when meeting real-world scenarios that never encountered before. To conquer it, we develop a novel bilevel learning scheme for fast adaptation to bridge the gap between low-light scenes. Concretely, we construct a Retinex-induced encoder-decoder with an adaptive denoising mechanism, aiming at covering more practical cases. Different from existing works that directly learn model parameters by using the massive data, we provide a new hyperparameter optimization perspective to formulate a bilevel learning scheme towards general low-light scenarios. This scheme depicts the latent correspondence (i.e., scene-irrelevant encoder) and the respective characteristic (i.e., scene-specific decoder) among different data distributions. Due to the expensive inner optimization, estimating the hyper-parameter gradient exactly can be prohibitive, we develop an approximate hyper-parameter gradient method by introducing the one-step forward approximation and finite difference approximation to ensure the high-efficient inference. Extensive experiments are conducted to reveal our superiority against other state-of-the-art methods. A series of analytical experiments are also executed to verify our effectiveness.

Recent deep networks have convincingly demonstrated high capability in crowd counting, which is a critical task attracting widespread attention due to its various industrial applications. Despite such progress, trained data-dependent models usually can not generalize well to unseen scenarios because of the inherent domain shift. To facilitate this issue, this paper proposes a novel adversarial scoring network (ASNet) to gradually bridge the gap across domains from coarse to fine granularity. In specific, at the coarse-grained stage, we design a dual-discriminator strategy to adapt source domain to be close to the targets from the perspectives of both global and local feature space via adversarial learning. The distributions between two domains can thus be aligned roughly. At the fine-grained stage, we explore the transferability of source characteristics by scoring how similar the source samples are to target ones from multiple levels based on generative probability derived from coarse stage. Guided by these hierarchical scores, the transferable source features are properly selected to enhance the knowledge transfer during the adaptation process. With the coarse-to-fine design, the generalization bottleneck induced from the domain discrepancy can be effectively alleviated. Three sets of migration experiments show that the proposed methods achieve state-of-the-art counting performance compared with major unsupervised methods., Accepted by ACMMM2021

Traditional learning systems are trained in closed-world for a fixed number of classes, and need pre-collected datasets in advance. However, new classes often emerge in real-world applications and should be learned incrementally. For example, in electronic commerce, new types of products appear daily, and in a social media community, new topics emerge frequently. Under such circumstances, incremental models should learn several new classes at a time without forgetting. We find a strong correlation between old and new classes in incremental learning, which can be applied to relate and facilitate different learning stages mutually. As a result, we propose CO-transport for class Incremental Learning (COIL), which learns to relate across incremental tasks with the class-wise semantic relationship. In detail, co-transport has two aspects: prospective transport tries to augment the old classifier with optimal transported knowledge as fast model adaptation. Retrospective transport aims to transport new class classifiers backward as old ones to overcome forgetting. With these transports, COIL efficiently adapts to new tasks, and stably resists forgetting. Experiments on benchmark and real-world multimedia datasets validate the effectiveness of our proposed method., Accepted to ACM Multimedia 2021

The Hexad user types model is often used in the gamification community to tailor gamified systems. However, most often, it requires users to fill out a questionnaire, preventing an automated adaptation of the interactive system. For this reason, we explored the potential of using mobile banking data to automate the profiling of Hexad user types. In our study, we conducted an expert consensus study to research whether a group of experts (N=11) in the field of gamification and banking perceive there is a relation between the Hexad user types and banking data. The results show that experts find this relation present, indicating that automating the calculation of Hexad user types from banking data could be feasible.

The COVID-19 pandemic brought harsh social restrictions as part of the strategy to deal with the situation, which also required innovation to quickly adapt and keep providing substitutes through digital solutions. In response to this emergency, the local entrepreneurship ecosystem of Sao Joao del-Rei, Brazil, spent a great effort to tackle this new reality. This article presents the experience of performing virtual events to foster innovation and raise awareness of the local entrepreneurship culture. The first event was developed under the hackathon format, aiming to encourage the local community to propose technological innovation by developing solutions to assist distance learning in universities and high schools. The other event was a virtual fair presenting local startups to the audience, besides providing online courses related to entrepreneurship and new digital technologies. It is worth noting that those events were organized by a strategic coalition following the triple helix model of innovation, gathering universities, industry, and government. The results achieved through those events reinforce the importance of having an organized local strategic coalition to provide a fast response to social emergencies and provide better adaptation capability to society by creating innovations based on digital entrepreneurship.

We present an optimization-based approach that automatically adapts Mixed Reality (MR) interfaces to different physical environments. Current MR layouts, including the position and scale of virtual interface elements, need to be manually adapted by users whenever they move between environments, and whenever they switch tasks. This process is tedious and time consuming, and arguably needs to be automated for MR systems to be beneficial for end users. We contribute an approach that formulates this challenge as a combinatorial optimization problem and automatically decides the placement of virtual interface elements in new environments. To achieve this, we exploit the semantic association between the virtual interface elements and physical objects in an environment. Our optimization furthermore considers the utility of elements for users’ current task, layout factors, and spatio-temporal consistency to previous layouts. All those factors are combined in a single linear program, which is used to adapt the layout of MR interfaces in real time. We demonstrate a set of application scenarios, showcasing the versatility and applicability of our approach. Finally, we show that compared to a naive adaptive baseline approach that does not take semantic associations into account, our approach decreased the number of manual interface adaptations by 33%.

We demonstrate the design of adaptive tools for motor skill training that use shape change to automatically vary task difficulty based on a learner’s performance. Studies [1] have shown that automatically-adaptive tools lead to significantly higher learning gains when compared to non-adaptive and manually-adaptive tools. We demonstrate the use of Adapt2Learn [2] - a toolkit that supports designers in building adaptive training tools. Adapt2Learn auto-generates an algorithm that converts a learner’s performance data into adaptation states during motor skill training. This algorithm, that maintains the training difficulty at the ’optimal challenge point’, can be uploaded to the micro-controller to convert several shape-changing tools into adaptive tools for motor skill training. We demonstrate 7 prototypes of adaptive tools for motor-skill learning to show applications in sports, music, rehabilitation, and accessibility.

Managing dynamic variability has motivated several researchers to combine Dynamic Software Product Lines (DSPLs) practices with runtime variability mechanisms. By combining these approaches, a DSPL acquires important features, ranging from the ability to reconfigure by changing the context, adding or removing features, crash recovery, and re-adaptation based on changes in the model’s features. Feature model (FM) is an important artifact of a DPSL and there is a lack of tools that support the modeling of this artifact. We have extended the DyMMer tool for modeling FM of DSPLs from an adaptation mechanism based on MAPE-K to solve this problem. We migrated the DyMMer tool to a web version and incorporated new features: (i) modeling of FMs from SPLs and DSPLs, (ii) development of an adaptation mechanism for FM of DSPLs, (iii) repository of FMs, (iv) inclusion of thresholds for measures, and (v) user authentication. We believe that this tool is useful for research in the area of DSPLs, and also for dynamic domain modeling and evaluation. Video: https://youtu.be/WVHW6bI8ois

Digital companions and conversational agents are becoming increasingly popular in our everyday lives. Natural language interfaces play an important role in ubiquitous computing: voice assistants are used to control smart home devices and smartphone applications; chatbots serve as an interface to solve tasks and acquire information easily. However, misunderstandings due to non-standard language, expressions that serve social functions without conveying information, or a lack of situational awareness still pose problems for these interfaces. Humans are able to prevent or repair communication failures by imitating their conversation partner’s lexical choices, sentence structures, and overall language style; a mechanism known as Linguistic Alignment. In this paper we present different strategies to easily integrate an alignment effect in natural language interfaces. We implemented a chatbot that imitates alignment and tested it in an online user study with 75 participants. Our results show that alignment helps to decrease user frustration and perceived task workload.

The development of architectural connectors as first-class software entities is a long-standing software engineering promise that has not been fully realized thus far. This observation is especially true in self-adaptive software systems, where most advances within academia and industry have targeted software components as the primary locus for run-time adaptation. In this paper, we revisit the evolutionary history of architectural connectors and discuss the challenges of implementing architectural connectors as first-class self-adaptation entities in the domain of modern microservice applications. We then make a case for our ongoing work on a cloud-native self-adaptive service mesh framework that builds on recent container orchestration, self-adaptation, and service mesh technologies.

The COVID-19 pandemic has lasted for over 1 year. During the pandemic, several software development companies migrated to working from home (WFH). Several studies have recorded an initial moment of adaptation to WFH. This study investigated how Brazilian software development teams dealt with WFH after 1 year of a pandemic and how this change in work was on the software development process. We applied a survey with 67 participants of software development teams and investigated aspects, as: work routine, collaboration, communication, productivity, financial assistance and software development process. We performed a quantitative and qualitative analysis of the survey results and compared them with previous studies. Our key findings are: (i) 56.71% of the participants said that it has an impact on the work routine related to work overload and exceeds the company’s standard hours; (ii) 92.54% of the participants consider their team to be collaborative; (iii) 82.70% are satisfied with the communication at WFH; (iv) 86.55% has meetings once or twice a day; (v) 59.7% had an improvement in productivity in WFH and 75% are satisfied with their productivity; (vi) 79.60% receive some assistance from the company for WFH; (vii) 82.1% said that the company intends or probably intends to continue in the WFH, and (viii) 55.2% said they had changed in the software development process due to WFH. The main positive changes in the process are related to the practices of: daily meetings, agile processes, code review, pair programming and code versioning.

Under the current background of discipline integration, the curriculum reform of interdisciplinary cooperative teaching is a required course in every school. This teaching mode of interdisciplinary cooperation roughly goes through four stages which are the stage of subject teaching and textbook adaptation, the stage of individualized education and interdisciplinary cooperation. In this paper, the author focuses on the development of interdisciplinary cooperative teaching mode under the background of discipline integration. This mode is a brand new teaching reform developed on the basis of the predecessor's individualized education plan mode, whose core is curriculum reform and professional team building; and this model provides relevant strategies and supporting routine activities and thematic teaching activities based on comprehensive evaluation, so that the teaching effectiveness of the school presents a spiraling trend with the steady advancement of the curriculum reform.

This paper shows the results of a study developed to analyze some of the existing technology acceptance models, with the goal identify which is the most appropriate and which elements should be included in a study of technological acceptance of games for older adults implemented on virtual assistants. For the analysis of the models, relevant studies that used them or did a comparative study were considered. The preliminary results make it possible to determine that special needs of older adults should be considered in terms of physical conditions and cognitive status regardless of the technology acceptance model used.

Some of the biggest challenges in applying Augmented Reality (AR) technologies to the industry floor are in the form factor, and safety requirements of the head worn display. This includes alleviating issues such as peripheral view occlusion, and adaptation to personal protective equipment. In this work we present the design of Project Ariel, an Open Source 3D printable display specifically designed for use in industrial environments. It is our hope that with this technology, the average tradesman can utilize the powerful visualizations AR has to offer, significantly improving their daily work flow.

Recommender systems evaluation has evolved rapidly in recent years. However, for offline evaluation, accuracy is the de facto standard for assessing the superiority of one method over another, with most research comparisons focused on tasks ranging from rating prediction to ranking metrics for top-n recommendation. Simultaneously, recommendation diversity and novelty have become recognized as critical to users’ perceived utility, with several new metrics recently proposed for evaluating these aspects of recommendation lists. Consequently, the accuracy-diversity dilemma frequently shows up as a choice to make when creating new recommendation algorithms. We propose a novel adaptation of a unified metric, derived from one commonly used for search system evaluation, to Recommender Systems. The proposed metric combines topical diversity and accuracy, and we show it to satisfy a set of desired properties that we formulate axiomatically. These axioms are defined as fundamental constraints that a good unified metric should always satisfy. Moreover, beyond the axiomatic analysis, we present an experimental evaluation of the metric with collaborative filtering data. Our analysis shows that the metric respects the desired theoretical constraints and behaves as expected when performing offline evaluation.

Human factors research and engineering of advanced driving assistance systems (ADAS) must consider how drivers adapt to their presence. The major obstruction to this at the moment is poor understanding of the details of the adaptive processes that the human cognition undergoes when faced with such changes. This paper presents a simulation model that predicts how drivers adapt to a steering assistance system. Our approach is based on computational rationality, and demonstrates how task interleaving strategies adapt to the task environment and the driver’s goals and cognitive limitations. A supervisor controls eye movements between the driving and non-driving tasks, making this choice on the basis of maximising expected joint task utility. The model predicts that with steering assistance, drivers’ in car glance durations increase. We also show that this adaptation leads to risky driving in cases where the reliability of the system is compromised.

6 páginas.- 2 figura.- 29 referencias.- 1st Conference on Information Technology for Social Good, GoodIT 2021, Rome 9-11 September 2021, New challenges such as climate change and sustainability arise in society influencing not only environmental issues but human's health directly. To face these new challenges IT technologies and their application to environmental intelligent monitoring become into a powerful tool to set new policies and blueprints to contribute to social good. In the new H2020 project, WatchPlant will provide new tools for environmental intelligence monitoring by the use of plants as "well-being"sensors of the environment they inhabit. This will be possible by equipping plants with a net of communicated wireless self-powered sensors, coupled with artificial intelligence (AI) to become plants into "biohybrid organisms"to test exposure-effects links between plant and the environment. It will become plants into a new tool to be aware of the environment status in a very early stage towards in-situ monitoring. Additionally, the system is devoted to be sustainable and energy-efficient thanks to the use of clean energy sources such as solar cells and a enzymatic biofuel cell (BFC) together with its self-deployment, self-awareness, adaptation, artificial evolution and the AI capabilities. In this concept paper, WatchPlant will envision how to face this challenge by joining interdisciplinary efforts to access the plant sap for energy harvesting and sensing purposes and become plants into "biohybrid organisms"to benefit social good in terms of environmental monitoring in urban scenarios. © 2021 Owner/Author., Project WatchPlant has received funding from the European Union’s Horizon 2020 research and innovation program under the FET grant agreement, no. 101017899

As automated vehicles become more prevalent, designing interfaces that best fit all users, especially ones in minority populations, is a pressing but difficult goal. System-driven adaptation is a commonly used approach as it is easier and created by experts but, has innate flaws. Customization, on the other hand, allows users to consciously alter the interface to appear and operate in a manner most suited to their needs and wants. However, various components of the interface have different constraints, capabilities, and requirements with the amount of customization appropriate. In this workshop, we will dissect an expansive taxonomy for customization and develop a series of levels in order to get the full benefits from customization, which in turn can help engineers and designers in creating more user-centered systems.

Ambient intelligence represents a paradigm in which the user does not react to the environment, but vice versa. Accordingly, smart environments can react to the presence and activities of users and support them unobtrusively from the background. Especially in the context of accessibility, this offers great potential that has so far only been demonstrated for individual user groups. To overcome this limitation, we propose the automated, user- and context-related adaptation of the modality as well as locality of the representation of building information in the form of both an adjustable table as well as two displays on the basis of a prototype for a library information center. For being independent from material and regulatory restrictions and for better planability (especially with the ongoing COVID-19 pandemic) we used in addition to the hardware components also a Virtual Reality simulation, which proved to be very useful. Further optimization and evaluation will be needed for a more in depth understanding and dissemination in the long run, yet our prototype aims to help fostering further activities in the field of ambient intelligence, accessibility and virtual reality as a planning tool.

Recent measurement studies show that commercial mmWave 5G can indeed offer ultra-high bandwidth (up to 2 Gbps), capable of supporting bandwidth-intensive applications such as ultra-HD (UHD) 4K/8K and volumetric video streaming on mobile devices. However, mmWave 5G also exhibits highly variable throughput performance and incurs frequent handoffs (e.g., between 5G and 4G), due to its directional nature, signal blockage and other environmental factors, especially when the device is mobile. All these issues make it difficult for applications to achieve high Quality of Experience (QoE). In this paper, we advance several new mechanisms to tackle the challenges facing UHD video streaming applications over 5G networks, thereby making them {\em 5G-aware}. We argue for the need to employ machine learning (ML) for effective throughput prediction to aid applications in intelligent bitrate adaptation. Furthermore, we advocate {\em adaptive content bursting}, and {\em dynamic radio (band) switching} to allow the 5G radio network to fully utilize the available radio resources under good channel/beam conditions, whereas dynamically switched radio channels/bands (e.g., from 5G high-band to low-band, or 5G to 4G) to maintain session connectivity and ensure a minimal bitrate. We conduct initial evaluation using real-world 5G throughput measurement traces. Our results show these mechanisms can help minimize, if not completely eliminate, video stalls, despite wildly varying 5G throughput.

Resource-driven systems have tasks that are bound by limited resources. These systems must adapt their tasks that cannot gain access to sufficient resources. This dissertation proposes a new resource-driven adaptation approach, which aims to support (1) task prioritisation using multiple criteria such as the time of day that a task should be executed, the role of involved users, and selection of the least costly adaptation types; (2) collaboration between a human user and a software tool for preparing adapted task behaviour to be used when resources are substituted; and (3) resource extensibility and heterogeneity. The proposed approach is being implemented and will be evaluated with scenarios from enterprise applications.

The process of digitalization of societies and innovation is involving the fast introduction of new technologies in different sectors. However, the development of technology represents a challenge as it involves technical, legal, economic and social aspects that have to be considered since its conception or design. The aim of this paper is to offer an adaptation of an existing legal methodology, the Data Protection Impact Assessment, as a legal obligation to evaluate technological proposals and assure compliance with privacy by design requirements. For that purpose, we will refer to the specific case of Identity Management technologies. We introduce the main challenges in the sector of Digital Identity Management as well as the importance of covering the “architecture” and “user” sides in the development of safer technologies by citing concrete examples. Finally, in order to provide a more practical view of the methodology to adapt the Data Protection Impact Assessment, we refer to the work developed in the research project OLYMPUS in the evaluation of its privacy implications. By introducing this example, the paper offers a specific methodology directly reusable for the study of technological proposals in IdM but that can be adapted to any other sector.

Federated learning learns from scattered data by fusing collaborative models from local nodes. However, the conventional coordinate-based model averaging by FedAvg ignored the random information encoded per parameter and may suffer from structural feature misalignment. In this work, we propose Fed2, a feature-aligned federated learning framework to resolve this issue by establishing a firm structure-feature alignment across the collaborative models. Fed2 is composed of two major designs: First, we design a feature-oriented model structure adaptation method to ensure explicit feature allocation in different neural network structures. Applying the structure adaptation to collaborative models, matchable structures with similar feature information can be initialized at the very early training stage. During the federated learning process, we then propose a feature paired averaging scheme to guarantee aligned feature distribution and maintain no feature fusion conflicts under either IID or non-IID scenarios. Eventually, Fed2 could effectively enhance the federated learning convergence performance under extensive homo- and heterogeneous settings, providing excellent convergence speed, accuracy, and computation/communication efficiency., Accepted in KDD 2021

Deep learning has shown powerful performances in many fields, however its black-box nature hinders its further applications. In response, explainable artificial intelligence emerges, aiming to explain the predictions and behaviors of deep learning models. Among many explanation methods, counterfactual explanation has been identified as one of the best methods due to its resemblance to human cognitive process: to deliver an explanation by constructing a contrastive situation so that human may interpret the underlying mechanism by cognitively demonstrating the difference. In this tutorial, we will introduce the cognitive concept and characteristics of counterfactual explanation, its computational form, mainstream methods, and various adaptation in terms of different explanation settings. In addition, we will demonstrate several typical use cases of counterfactual explanations in popular research areas. Finally, in light of practice, we outline the potential applications of counterfactual explanations like data augmentation or conversation system. We hope this tutorial can help the participants get an overview sense of counterfactual explanations.

Artificial Intelligence and machine learning techniques can offer powerful tools for addressing the greatest challenges facing humanity and helping society adapt to a rapidly changing climate, respond to disasters and pandemic crisis, and reach the United Nations (UN) Sustainable Development Goals (SDGs) by 2030. In recent approaches for mitigation and adaptation, data analytics and ML are only one part of the solution that requires interdisciplinary and methodological research and innovations. For example, challenges include multi-modal and multi-source data fusion to combine satellite imagery with other relevant data, handling noisy and missing ground data at various spatio-temporal scales, and ensembling multiple physical and ML models to improve prediction accuracy. Despite recognized successes, there are many areas where ML is not applicable, performs poorly or gives insights that are not actionable. This tutorial will survey the recent and significant contributions in KDD and ML for sustainable development and will highlight current challenges that need to be addressed to transform and equip engaged sustainability science with robust ML-based tools to support actionable decision-making for a more sustainable future.