Ubiquitous wearable and mobile devices provide access to a diverse set of data. However, the mobility demand for our devices naturally imposes constraints on their computational and communication capabilities. A solution is to locally learn knowledge from data captured by ubiquitous devices, rather than to store and transmit the data in its original form. In this paper, we develop a federated learning framework, called Centaur, to incorporate on-device data selection at the edge, which allows partition-based training of a deep neural nets through collaboration between constrained and resourceful devices within the multidevice ecosystem of the same user. We benchmark on five neural net architecture and six datasets that include image data and wearable sensor time series. On average, Centaur achieves ~19% higher classification accuracy and ~58% lower federated training latency, compared to the baseline. We also evaluate Centaur when dealing with imbalanced non-iid data, client participation heterogeneity, and different mobility patterns. To encourage further research in this area, we release our code at github.com/nokia-bell-labs/data-centric-federated-learning.

Self-supervised learning (SSL) has shown remarkable performance in computer vision tasks when trained offline. However, in a Continual Learning (CL) scenario where new data is introduced progressively, models still suffer from catastrophic forgetting. Retraining a model from scratch to adapt to newly generated data is time-consuming and inefficient. Previous approaches suggested re-purposing self-supervised objectives with knowledge distillation to mitigate forgetting across tasks, assuming that labels from all tasks are available during fine-tuning. In this paper, we generalize self-supervised continual learning in a practical setting where available labels can be leveraged in any step of the SSL process. With an increasing number of continual tasks, this offers more flexibility in the pre-training and fine-tuning phases. With Kaizen1, we introduce a training architecture that is able to mitigate catastrophic forgetting for both the feature extractor and classifier with a carefully designed loss function. By using a set of comprehensive evaluation metrics reflecting different aspects of continual learning, we demonstrated that Kaizen significantly outperforms previous SSL models in competitive vision benchmarks, with up to 16.5% accuracy improvement on split CIFAR-100. Kaizen is able to balance the trade-off between knowledge retention and learning from new data with an end-to-end model, paving the way for practical deployment of continual learning systems.

Limited availability of labeled data for machine learning on multimodal time-series extensively hampers progress in the field. Self-supervised learning (SSL) is a promising approach to learn data representations without relying on labels. However, existing SSL methods require expensive computations of negative pairs and are typically designed for single modalities, which limits their versatility. We introduce CroSSL (Cross-modal SSL), which puts forward two novel concepts: masking intermediate embeddings produced by modality-specific encoders, and their aggregation into a global embedding through a cross-modal aggregator CroSSL allows for handling missing modalities and end-to-end cross-modal earning without requiring prior data preprocessing for handling missing inputs or negative-pair sampling for contrastive learning. We evaluate our method on a wide range of data, including motion sensors such as accelerometers or gyroscopes and biosignals (heart rate, electroencephalograms, electromyograms, electrooculograms, and electrodermal). Overall, CroSSL outperforms previous SSL and supervised benchmarks using minimal labeled data, and also sheds light on how latent masking can improve cross-modal learning.

Despite impressive results, deep learning-based technologies also raise severe privacy and environmental concerns induced by the training procedure often conducted in data centers. In response, alternatives to centralized training such as Federated Learning (FL) have emerged. FL is now starting to be deployed at a global scale by companies that must adhere to new legal demands and policies originating from governments and social groups advocating for privacy protection. However, the potential environmental impact related to FL remains unclear and unexplored. This article offers the first-ever systematic study of the carbon footprint of FL. We propose a rigorous model to quantify the carbon footprint, hence facilitating the investigation of the relationship between FL design and carbon emissions. We also compare the carbon footprint of FL to traditional centralized learning. Our findings show that, depending on the configuration, FL can emit up to two orders of magnitude more carbon than centralized training. However, in certain settings, it can be comparable to centralized learning due to the reduced energy consumption of embedded devices. Finally, we highlight and connect the results to the future challenges and trends in FL to reduce its environmental impact, including algorithms efficiency, hardware capabilities, and stronger industry transparency.

We present SensiX++, a multi-tenant runtime for adaptive model execution with integrated MLOps on edge devices, e.g., a camera, a microphone, or IoT sensors. SensiX++ operates on two fundamental principles: highly modular componentisation to externalise data operations with clear abstractions and document-centric manifestation for system-wide orchestration. First, a data coordinator manages the lifecycle of sensors and serves models with correct data through automated transformations. Next, a resource-aware model server executes multiple models in isolation through model abstraction, pipeline automation, and feature sharing. An adaptive scheduler then orchestrates the best-effort executions of multiple models across heterogeneous accelerators, balancing latency and throughput. Finally, microservices with REST APIs serve synthesised model predictions, system statistics, and continuous deployment. Collectively, these components enable SensiX++ to serve multiple models efficiently with fine-grained control on edge devices while minimising data operation redundancy, managing data and device heterogeneity, and reducing resource contention. We benchmark SensiX++ with 10 different vision and acoustics models across various multi-tenant configurations on different edge accelerators (Jetson AGX and Coral TPU) designed for sensory devices. We report on the overall throughput and quantified benefits of various automation components of SensiX++ and demonstrate its efficacy in significantly reducing operational complexity and lowering the effort to deploy, upgrade, reconfigure, and serve embedded models on edge devices.

Multiple sensory devices on and around us are on the rise and require us to redesign a system to make an inference of ML models accurate, robust, and efficient at the deployment time. While this multiplicity opens up an exciting opportunity to leverage sensor redundancy, it is still extremely challenging to benefit from such multiplicity and boost the runtime performance of deployed ML models without model retraining and engineering. From our experience, we uncovered two prime caveats, device and data variabilities, that affect the runtime performance of ML models. We develop an ML system that addresses these variabilities without modifying deployed models by building on prior algorithmic work. It decouples model execution from sensor data and employs two essential operations between them: a) device-to-device data translation for principled mapping of training and inference data and b) quality-aware dynamic selection of the execution pipeline as a function of runtime accuracy. We evaluate the system on wearable devices with motion and audio-based models. The results show that ML models achieve a 7-13% increase in runtime accuracy solely by running on our system, and the increase goes up to 30% in dynamic environments, at the expense of 3 mW on the host device.

Billions of distributed, heterogeneous, and resource constrained IoT devices deploy on-device machine learning (ML) for private, fast, and offline inference on personal data. On-device ML is highly context dependent and sensitive to user, usage, hardware, and environment attributes. This sensitivity and the propensity toward bias in ML makes it important to study bias in on-device settings. Our study is one of the first investigations of bias in this emerging domain and lays important foundations for building fairer on-device ML. We apply a software engineering lens, investigating the propagation of bias through design choices in on-device ML workflows. We first identifyreliability biasas a source of unfairness and propose a measure to quantify it. We then conduct empirical experiments for a keyword spotting task to show how complex and interacting technical design choices amplify and propagatereliability bias. Our results validate that design choices made during model training, like the sample rate and input feature type, and choices made to optimize models, like light-weight architectures, the pruning learning rate, and pruning sparsity, can result in disparate predictive performance across male and female groups. Based on our findings, we suggest low effort strategies for engineers to mitigate bias in on-device ML.

Sleep deficiency and disorders are one of the most unsolved public health challenges of modern times. Music therapy is a promising approach, offering a cheap and non-invasive solution to improve sleep quality. However, the choice of therapeutic sleep music is highly limited for users because such music needs to be specially chosen and made by sleep therapists. It could potentially lead to the inefficiency of music therapy if users get bored after listening to the same set of music repeatedly. In this paper, we take the first step towards generating personalized sleep therapy music. Firstly, through an in-depth feature analysis, we investigate the importance of various musical and acoustic features of therapy music. Grounded on our findings, we design a style transfer framework called SleepGAN which induces therapeutic features into music from different genres. We show that, compared to baselines, the music generated by SleepGAN has a higher similarity to the sleep music designed by experts.

Federated learning is generally used in tasks where labels are readily available (e.g., next word prediction). Relaxing this constraint requires design of unsupervised learning techniques that can support desirable properties for federated training: robustness to statistical/systems heterogeneity, scalability with number of participants, and communication efficiency. Prior work on this topic has focused on directly extending centralized self-supervised learning techniques, which are not designed to have the properties listed above. To address this situation, we propose Orchestra, a novel unsupervised federated learning technique that exploits the federation’s hierarchy to orchestrate a distributed clustering task and enforce a globally consistent partitioning of clients’ data into discriminable clusters. We show the algorithmic pipeline in Orchestra guarantees good generalization performance under a linear probe, allowing it to outperform alternative techniques in a broad range of conditions, including variation in heterogeneity, number of clients, participation ratio, and local epochs.

Breakthroughs in unsupervised domain adaptation (uDA) can help in adapting models from a label-rich source domain to unlabeled target domains. Despite these advancements, there is a lack of research on how uDA algorithms, particularly those based on adversarial learning, can work in distributed settings. In real-world applications, target domains are often distributed across thousands of devices, and existing adversarial uDA algorithms – which are centralized in nature – cannot be applied in these settings. To solve this important problem, we introduce FRuDA: an end-to-end framework for distributed adversarial uDA. Through a careful analysis of the uDA literature, we identify the design goals for a distributed uDA system and propose two novel algorithms to increase adaptation accuracy and training efficiency of adversarial uDA in distributed settings. Our evaluation of FRuDA with five image and speech datasets show that it can boost target domain accuracy by up to 50% and improve the training efficiency of adversarial uDA by at least$11\times$.

We are witnessing a trend of users owning multiple data-generating wearable and IoT devices that continuously capture sensor data pertaining to a user's activities and context. Federated Learning is a potential technique to derive meaningful insights from this sensor data in a privacy-preserving way without revealing the raw sensor data to a central server. In this paper, we introduce a new problem setting in this multi-device context called Federated Learning in Multi-Device Local Networks (FL-MDLN). We identify core challenges for FL-MDLN in relation to its federation architecture, and statistical and systems heterogeneity across multiple users and multiple devices. Then, we introduce a new user-as-client (UAC) federation architecture, and propose various device selection strategies to counter statistical and systems heterogeneity in FL-MDLN. Early empirical findings show that our proposed techniques improve model test accuracy as well as battery power efficiency in FL. Based on these findings, we elucidate open research questions and future work in FL-MDLN.

When deploying machine learning (ML) models on embedded and IoT devices, performance encompasses more than an accuracy metric: inference latency, energy consumption, and model fairness are necessary to ensure reliable performance under heterogeneous and resource-constrained operating conditions. To this end, prior research has studied model-centric approaches, such as tuning the hyperparameters of the model during training and later applying model compression techniques to tailor the model to the resource needs of an embedded device. In this paper, we take a data-centric view of embedded ML and study the role that pre-processing parameters in the data pipeline can play in balancing the various performance metrics of an embedded ML system. Through an in-depth case study with audio-based keyword spotting (KWS) models, we show that pre-processing parameter tuning is a remarkable tool that model developers can adopt to trade-off between a model's accuracy, fairness, and system efficiency, as well as to make an embedded ML model resilient to unseen deployment conditions.

In chronic pain rehabilitation, physiotherapists adapt physical activity to patients’ performance based on their expression of protective behavior, gradually exposing them to feared but harmless and essential everyday activities. As rehabilitation moves outside the clinic, technology should automatically detect such behavior to provide similar support. Previous works have shown the feasibility of automatic protective behavior detection (PBD) within a specific activity. In this article, we investigate the use of deep learning for PBD across activity types, using wearable motion capture and surface electromyography data collected from healthy participants and people with chronic pain. We approach the problem by continuously detecting protective behavior within an activity rather than estimating its overall presence. The best performance reaches mean F1 score of 0.82 with leave-one-subject-out cross validation. When protective behavior is modeled per activity type, performance achieves a mean F1 score of 0.77 for bend-down, 0.81 for one-leg-stand, 0.72 for sit-to-stand, 0.83 for stand-to-sit, and 0.67 for reach-forward. This performance reaches excellent level of agreement with the average experts’ rating performance suggesting potential for personalized chronic pain management at home. We analyze various parameters characterizing our approach to understand how the results could generalize to other PBD datasets and different levels of ground truth granularity.

We are witnessing unprecedented datafication of the society we live in, alongside rapid advances in the fields of Artificial Intelligence and Machine Learning. However, emergent data-driven applications are systematically discriminating against many diverse populations. A major driver of the bias are the data, which typically align with predominantly Western definitions and lack representation from multilingually diverse and resource-constrained regions across the world. Therefore, data-driven approaches can benefit from integration of a more human-centred orientation before being used to inform the design, deployment, and evaluation of technologies in various contexts. This workshop seeks to advance these and similar conversations, by inviting researchers and practitioners in interdisciplinary domains to engage in conversation around how appropriate human-centred design can contribute to addressing data-related challenges among marginalised and under-represented/underserved groups.

This paper introduces a new dataset, Libri-Adapt, to support unsupervised domain adaptation research on speech recognition models. Built on top of the LibriSpeech corpus, Libri-Adapt contains 7200 hours of English speech recorded on mobile and embedded-scale microphones, and spans 72 different domains that are representative of the challenging practical scenarios encountered by ASR models. More specifically, Libri-Adapt facilitates the study of domain shifts in ASR models caused by a) different acoustic environments, b) variations in speaker accents, c) previously unexplored factors such as heterogeneity in the hardware and platform software of the microphones, and d) a combination of the aforementioned three shifts. We also provide a number of baseline results quantifying the impact of these domain shifts on the Mozilla DeepSpeech2 ASR model.

Unsupervised domain adaptation using adversarial learning has shown promise in adapting speech models from a labeled source domain to an unlabeled target domain. However, prior works make a strong assumption that the label spaces of source and target domains are identical, which can be easily violated in real-world conditions. We present AMLS, an end-to-end architecture that performs Adaptation under Mismatched Label Spaces using two weighting schemes to separate shared and private classes in each domain. An evaluation on three speech adaptation tasks, namely gender, microphone, and emotion adaptation, shows that AMLS provides significant accuracy gains over baselines used in speech and vision adaptation tasks. Our contribution paves the way for applying UDA to speech models in unconstrained settings with no assumptions on the source and target label spaces.

Conversational agents are increasingly becoming digital partners of our everyday computing experiences offering a variety of purposeful information and utility services. Although rich on competency, these agents are entirely oblivious to their users' situational and emotional context today and incapable of adjusting their interaction style and tone contextually. To this end, we present a mixed-method study that informs the design of a situation- and emotion-aware conversational agent for kinetic earables. We surveyed 280 users, and qualitatively interviewed 12 users to understand their expectation from a conversational agent in adapting the interaction style. Grounded on our findings, we develop a first-of-its-kind emotion regulator for a conversational agent on kinetic earable that dynamically adjusts its conversation style, tone, volume in response to users emotional, environmental, social and activity context gathered through speech prosody, motion signals and ambient sound. We describe these context models, the end-to-end system including a purpose-built kinetic earable and their real-world assessment. The experimental results demonstrate that our regulation mechanism invariably elicits better and affective user experience in comparison to baseline conditions in different real-world settings.

We explore a new variability observed in motion signals acquired from modern wearables. Wearing variability refers to the variations of the device orientation and placement across wearing events. We collect the accelerometer data on a smartwatch and an earbud and analyse how motion signals change due to the wearing variability. Our analysis shows that the wearing variability can bring an unexpected change to motion signals, not only from different users but also from different wearing sessions of the same user. We also provide empirical ranges of changes in device orientations.

In chronic pain physical rehabilitation, physiotherapists adapt exercise sessions according to the movement behavior of patients. As rehabilitation moves beyond clinical sessions, technology is needed to similarly assess movement behaviors and provide such personalized support. In this paper, as a first step, we investigate automatic detection of protective behavior (movement behavior due to pain-related fear or pain) based on wearable motion capture and electromyography sensor data. We investigate two recurrent networks (RNN) referred to as stacked-LSTM and dual-stream LSTM, which we compare with related deep learning (DL) architectures. We further explore data augmentation techniques and additionally analyze the impact of segmentation window lengths on detection performance. The leading performance of 0.815 mean F1 score achieved by stacked-LSTM provides important grounding for the development of wearable technology to support chronic pain physical rehabilitation during daily activities.

Deep learning has enabled personal and IoT devices to rethink microphones as a multi-purpose sensor for understanding conversation and the surrounding environment. This resulted in a proliferation of Voice Controllable Systems (VCS) around us. The increasing popularity of such systems is also prone to attracting miscreants, who often want to take advantage of the VCS without the knowledge of the user. Consequently, understanding the robustness of VCS, especially under adversarial attacks, has become an important research topic. Although there exists some previous work on audio adversarial attacks, their scopes are limited to embedding the attacks onto pre-recorded music clips, which when played through speakers cause VCS to misbehave. As an attack-audio needs to be played, the occurrence of this type of attacks can be suspected by a human listener. In this paper, we focus on audio-based Denial-of-Service (DoS) attack, which is unexplored in the literature. Contrary to previous work, we show that adversarial audio attacks in real-time and overthe-air are possible, while a user interacts with VCS. We show that the attacks are effective regardless of the user's command and interaction timings. In this paper, we present a first-of-itskind imperceptible and always-on universal audio perturbation technique that enables such DoS attack to be successful. We thoroughly evaluate the performance of the attacking scheme across (i) two learning tasks, (ii) two model architectures and (iii) three datasets. We demonstrate that the attack can introduce as high as 78% error rate in audio recognition tasks.

Unsupervised domain adaptation is emerging as a powerful technique to improve the generalizability of deep learning models to new image domains without using any labeled data in the target domain. In the literature, solutions which perform cross-domain feature-matching (e.g., ADDA), pixel-matching (CycleGAN), and combination of the two (e.g., CyCADA) have been proposed for unsupervised domain adaptation. Many of these approaches make a strong assumption that the source and target label spaces are the same, however in the real-world, this assumption does not hold true. In this paper, we propose a novel solution, FlexAdapt, which extends the state-of-the-art unsupervised domain adaptation approach of CyCADA to scenarios where the label spaces in source and target domains are only partially overlapped. Our solution beats a number of state-of-the-art baseline approaches by as much as 29% in some scenarios, and represent a way forward for applying domain adaptation techniques in the real world.

Mobile and embedded devices are increasingly using microphones and audio-based computational models to infer user context. A major challenge in building systems that combine audio models with commodity microphones is to guarantee their accuracy and robustness in the real-world. Besides many environmental dynamics, a primary factor that impacts the robustness of audio models is microphone variability. In this work, we propose Mic2Mic - a machine-learned system component - which resides in the inference pipeline of audio models and at real-time reduces the variability in audio data caused by microphone-specific factors. Two key considerations for the design of Mic2Mic were: a) to decouple the problem of microphone variability from the audio task, and b) put minimal burden on end-users to provide training data. With these in mind, we apply the principles of cycle-consistent generative adversarial networks (CycleGANs) to learn Mic2Mic using unlabeled and unpaired data collected from different microphones. Our experiments show that Mic2Mic can recover between 66% to 89% of the accuracy lost due to microphone variability for two common audio tasks.

Conversational agents are increasingly becoming digital partners of our everyday computing experiences offering a variety of purposeful information and utility services. Although rich on competency, these agents are entirely oblivious to their users' situational and emotional context today and incapable of adjusting their interaction style and tone contextually. To this end, we present a first-of-its-kind situation-aware conversational agent on kinetic earable that dynamically adjusts its conversation style, tone, volume in response to users emotional, environmental, social and activity context gathered through speech prosody, ambient sound and motion signatures.

The increasing availability of multiple sensory devices on or near a human body has opened brand new opportunities to leverage redundant sensory signals for powerful sensing applications. For instance, personal-scale sensory inferences with motion and audio signals can be done individually on a smartphone, a smartwatch, and even an earbud - each offering unique sensor quality, model accuracy, and runtime behaviour. At execution time, however, it is incredibly challenging to assess these characteristics to select the best device for accurate and resource-efficient inferences. To this end, we look at a quality-aware collaborative sensing system that actively interplays across multiple devices and respective sensing models. It dynamically selects the best device as a function of model accuracy at any given context. We propose two complementary techniques for the runtime quality assessment. Borrowing principles from active learning, our first technique runs on three heuristic-based quality assessment functions that employ confidence, margin sampling, and entropy of models' output. Our second technique is built with a siamese neural network and acts on the premise that runtime sensing quality can be learned from historical data. Our evaluation across multiple motion and audio datasets shows that our techniques provide 12% increase in overall accuracy through dynamic device selection at the average expense of 13 mW power on each device as compared to traditional single-device approaches.

A small variation in mobile hardware and software can potentially cause a significant heterogeneity or variation in the sensor data each device collects. For example, the microphone and accelerometer sensors on different devices can respond very differently to the same audio or motion phenomena. Other factors, like the instantaneous computational load on a smartphone, can cause key behavior like sensor sampling rates to fluctuate, further polluting the data. When sensing devices are deployed in unconstrained and real-world conditions, examples of sharply lower classification accuracy are observed due to what is collectively known as the sensing system heterogeneity. In this work, we take an unconventional approach and argue against solving individual forms of heterogeneity, e.g., improving OS behavior, or the quality/uniformity of components. Instead, we propose and build classifiers that themselves are more tolerant of these variations by leveraging deep learning and a data-augmented training process. Neither augmentation nor deep learning has previously been attempted to cope with sensor heterogeneity. We systematically investigate how these two machine learning methodologies can be adapted to solve such problems, and identify when and where they are able to be successful. We find that our proposed approach is able to reduce classifier errors on an average by 9% and 17% for a range of inertial-and audio-based mobile classification tasks.

No abstract available.

No abstract available.

We present eSense - an open and multi-sensory in-ear wearable platform for personal-scale behaviour analytics. eSense is a true wireless stereo (TWS) earbud and supports dual-mode Bluetooth and Bluetooth Low Energy. It is also augmented with a 6-axis in-ertial measurement unit and a microphone. We demonstrate the eSense platform, the data exploration tool with the open APIs for the real-time visualisation of multi-modal sensory data, and its manifestation in a 360° workplace well-being application.

Wearable devices with built-in cameras present interesting opportunities for users to capture various aspects of their daily life and are potentially also useful in supporting users with low vision in their everyday tasks. However, state-of-the-art image wearables available in the market are limited to capturing images periodically and do not provide any real-time analysis of the data that might be useful for the wearers. In this paper, we present DeepEye - a match-box sized wearable camera that is capable of running multiple cloud-scale deep learn- ing models locally on the device, thereby enabling rich analysis of the captured images in near real-time without offloading them to the cloud. DeepEye is powered by a commodity wearable processor (Snapdragon 410) which ensures its wearable form factor. The software architecture for DeepEye addresses a key limitation with executing multiple deep learning models on constrained hardware, that is their limited runtime memory. We propose a novel inference software pipeline that targets the local execution of multiple deep vision models (specifically, CNNs) by interleaving the execution of computation-heavy convolutional layers with the loading of memory-heavy fully-connected layers. Beyond this core idea, the execution framework incorporates: a memory caching scheme and a selective use of model compression techniques that further minimizes memory bottlenecks. Through a series of experiments, we show that our execution framework outperforms the baseline approaches significantly in terms of inference latency, memory requirements and energy consumption.

In this paper, we examine the potential of using mobile context to model user engagement. Taking an experimental approach, we systematically explore the dynamics of user engagement with a smartphone through three different studies. Specifically, to understand the feasibility of detecting user engagement from mobile context, we first assess an EEG artifact with 10 users and observe a strong correlation between automatically detected engagement scores and user's subjective perception of engagement. Grounded on this result, we model a set of application level features derived from smartphone usage of 10 users to detect engagement of a usage session using a Random Forest classifier. Finally, we apply this model to train a variety of contextual factors acquired from smartphone usage logs of 130 users to predict user engagement using an SVM classifier with a F1-Score of 0.82. Our experimental results highlight the potential of mobile contexts in designing engagement-aware applications and provide guidance to future explorations.

Face-to-face interactions have proven to accelerate team and larger organisation success. Many past research has explored the benefits of quantifying face-to-face interactions for informed workplace management, however to date, little attention has been paid to understand how the feedback on interaction behaviour is perceived at a personal scale. In this paper, we offer a reflection on the automated feedback of personal interactions in a workplace through a longitudinal study. We designed and developed a mobile system that captured, modelled, quantified and visualised face-to-face interactions of 47 employees for 4 months in an industrial research lab in Europe. Then we conducted semi-structured interviews with 20 employees to understand their perception and experience with the system. Our findings suggest that the short-term feedback on personal face-to-face interactions was not perceived as an effective external cue to promote self-reflection and that employees desire long-term feedback annotated with actionable attributes. Our findings provide a set of implications for the designers of future workplace technology and also opens up avenues for future HCI research on promoting self-reflection among employees.

We offer a reflection on the technology usage for workplace quantification through an in the wild study. Using a prototype Quantified Workplace system equipped with passive and participatory sensing modalities, we collected and visualized different workplace metrics (noise, color, air quality, self reported mood, and self reported activity) in two European offices of a research organization for a period of 4 months. Next we surveyed 70 employees to understand their engagement experience with the system. We then conducted semi-structured interviews with 20 employees in which they explained which workplace metrics are useful and why, how they engage with the system and what privacy concerns they have. Our findings suggest that sense of inclusion acts as the initial incentive for engagement which gradually translates into a habitual routine. We found that incorporation of an anonymous participatory sensing aspect into the system could lead to sustained user engagement. Compared to past studies we observed a shift in the privacy concerns, due to the trust and transparency of our prototype system. We conclude by providing a set of design principles for building future Quantified Workplace systems.

In this paper, we present the lessons learned by bringing in content and content-creators from local community media initiatives into the technological fold of the web. Our work focuses on the Community Radio (CR) ecosystem in India, and through extensive field-studies we develop an in-depth understanding of the operations, strengths and challenges of a CR station. Based on this, we outline the design of a system that combines the content creation processes of a CR station with the mobile web. The system was evaluated with the users of a CR station in Bangalore over a month long deployment. Our key take-away is that the incorporation of a mobile web based delivery system can play a critical role in expanding the reach and consumption of community media in the target communities. Conversely, the relevance of CR content and role of radio jockeys (as trusted members of the community) can be a key driver in adoption of the mobile web in these communities. Together, such a hybrid approach points the way forward for more successful deployments of community media systems, and reveals several interesting HCI issues to be studied further.

We explore the potential of using camera phones and pico projectors in rapid creation and presentation of digital content in a development context. A camera phone based content authoring application was designed and deployed with three different user populations in the domains of classroom education and health care. Our findings show that despite the variations in education levels, cultural background, and technology exposure, users successfully created and presented different forms of digital content using the camera phone and pico projector.

The Multiple Mice project demonstrated the financial [1] and learning [2] benefits of enabling students in resource-constrained schools to share one computer. In India, the lack of Multiple Mice authoring tools coupled with teachers' low computer proficiency means little or no customized content is created. This is problematic as the capability to create digital content enables teachers to prepare digital lessons that address the particular learning needs of their students [3], [4], [5].

Low educational levels hinder economic empowerment in developing countries. We make the case that educational games can impact children in the developing world. We report on exploratory studies with three communities in North and South India to show some problems with digital games that fail to match rural children's understanding of games, to highlight that there is much for us to learn about designing games that are culturally meaningful to them. We describe 28 traditional village games that they play, based on our contextual interviews. We analyze the mechanics in these games and compare these mechanics against existing videogames to show what makes traditional games unique. Our analysis has helped us to interpret the playability issues that we observed in our exploratory studies, and informed the design of a new videogame that rural children found to be more intuitive and engaging.

Literacy is one of the great challenges in the developing world. But universal education is an unattainable dream for those children who lack access to quality educational resources such as well-prepared teachers and schools. Worse, many of them do not attend school regularly due to their need to work for the family in the agricultural fields or households. This work commitment puts formal education far out of their reach. On the other hand, educational games on cellphones hold the promise of making learning more accessible and enjoyable. In our projects 4th year, we reached a stage where we could implement a semester-long pilot on cellphone-based learning. The pilot study took the form of an after-school program in a village in India. This paper reports on this summative learning assessment. While we found learning benefits across the board, it seemed that more of the gains accrued to those children who were better equipped to take advantage of this opportunity. We conclude with future directions for designing educational games that target less well-prepared children in developing regions.

Poor literacy remains a barrier to economic empowerment in the developing world. Of particular importance is fluency in a widely spoken "world language" such as English, which is typically a second language for these low-income learners. We make the case that mobile games on cellphones is an appropriate solution in the typical ecologies of developing regions. The challenge is to design e-learning games that are both educational and pleasurable for our target learners, who have limited familiarity with high technology. We propose the receptive-practice-activation cycle that could be used as the conceptual model for the designs. We then report how this format could be refined, based on our experiences in the field with three games that have collectively undergone nine rounds of iterations. In particular, it appears that maintaining a distinction between learning and fun to some extent is necessary for effective designs.

Low levels of education remain a barrier to economic empowerment in the developing world. In our work on English language learning among underserved communities in India since 2004, we have observed differences between school communities in terms of their access to educational opportunities outside school, access to ICTs including cellphones and digital gaming, enthusiasm for visitors, and the relationships between students. We report on these observations and argue that they call for the greater use of a comparative approach in constructing models of the micro-cultures at various schools, so that user-centered design processes and methods can better account for the unique differences across communities.