Automatic medical report generation from chest X-ray images is one possibility for assisting doctors to reduce their workload. However, the different patterns and data distribution of normal and abnormal cases can bias machine learning models. Previous attempts did not focus on isolating the generation of the abnormal and normal sentences in order to increase the variability of generated paragraphs. To address this, we propose to separate abnormal and normal sentence generation by using a dual word LSTM in a hierarchical LSTM model. In addition, we conduct an analysis on the distinctiveness of generated sentences compared to the BLEU score, which increases when less distinct reports are generated. Together with this analysis, we propose a way of selecting a model that generates more distinctive sentences. We hope our findings will help to encourage the development of new metrics to better verify methods of automatic medical report generation.

Twitter and other social media platforms are often used for sharing interest in products. The identification of purchase decision stages, such as in the AIDA model (Awareness, Interest, Desire, Action), can enable more personalized e-commerce services and a finer-grained targeting of ads than predicting purchase intent only. In this paper, we propose and analyze neural models for identifying the purchase stage of single tweets in a user's tweet sequence. In particular, we identify three challenges of purchase stage identification: imbalanced label distribution with a high number of negative instances, limited amount of training data, and domain adaptation with no or only little target domain data. Our experiments reveal that the imbalanced label distribution is the main challenge for our models. We address it with ranking loss and perform detailed investigations of the performance of our models on the different output classes. In order to improve the generalization of the models and augment the limited amount of training data, we examine the use of sentiment analysis as a complementary, secondary task in a multitask framework. For applying our models to tweets from another product domain, we consider two scenarios: For the first scenario without any labeled data in the target product domain, we show that learning domain-invariant representations with adversarial training is most promising while for the second scenario with a small number of labeled target examples, finetuning the source model weights performs best. Finally, we conduct several analyses, including extracting attention weights and representative phrases for the different purchase stages. The results suggest that the model is learning features indicative of purchase stages and that the confusion errors are sensible.

Human activity forecasting from videos in routine-based tasks is an open research problem that has numerous applications in robotics, visual monitoring and skill assessment. Currently, many challenges exist in activity forecasting because human actions are not fully observable from continuous recording. Additionally, a large number of human activities involve fine-grained articulated human motions that are hard to capture using frame-level representations. To overcome thesechallenges, we propose a method that forecasts human actions by learning the dynamics of local motion patterns extracted from dense trajectories using longshort-term memory (LSTM). The experiments on a pub-lic dataset validated the effectiveness of our proposed method in activity forecasting and demonstrate large improvements over the baseline two stream end-to-endmodel. We also learnt that human activity forecasting benefits from learning both the short-range motion pat-terns and long-term dependencies between actions.

A common issue in training a deep learning, abstractive summarization model is lack of a large set of training summaries. This paper examines techniques for adapting from a labeled source domain to an unlabeled target domain in the context of an encoder-decoder model for text generation. In addition to adversarial domain adaptation (ADA), we introduce the use of artificial titles and sequential training to capture the grammatical style of the unlabeled target domain. Evaluation on adapting to/from news articles and Stack Exchange posts indicates that the use of these techniques can boost performance for both unsupervised adaptation as well as fine-tuning with limited target data.

An open challenge in current telecommunication systems including Skype and other existing research systems is a lack of physical interaction, and consequently a restricted feeling of connection for users. For example, those telecommunication systems cannot allow remote users to move pieces of a board game while playing with a local user. We propose that installing a robot arm and teleoperating it can address the problem by enabling remote physical interaction. We compare three methods for remote control to study the relationship between connection, and how it relates to agency and autonomy for each control scheme.

Manufacturing environments require changes in work procedures and settings based on changes in product demand affecting the types of products for production. Resource re-organization and time needed for worker adaptation to such frequent changes can be expensive. For example, for each change, managers in a factory may be required to manually create a list of inventory items to be picked up by workers. Uncertainty in predicting the appropriate pick-up time due to differences in worker-determined routes may make it difficult for managers to generate a fixed schedule for delivery to the assembly line. To address these problems, we propose OPaPi, a human-centric system that improves the efficiency of manufacturing by optimizing parts pick-up routes and schedules. OPaPi leverages frequent pattern mining and the traveling salesman problem solver to suggest rack placement for more efficient routes. The system further employs interactive visualization to incorporate an expert’s domain knowledge and different manufacturing constraints for real-time adaptive decision making.

As our landscape of wearable technologies proliferates, we find more devices situated on our heads. However, many challenges hinder them from widespread adoption---from their awkward, bulky form factor (today's AR and VR goggles) to their socially stigmatized designs (Google Glass) and a lack of a well-developed head-based interaction design language. In this paper, we explore a socially acceptable, large, head-worn interactive wearable---a hat. We report results from a gesture elicitation study with 17 participants, extract a taxonomy of gestures, and define a set of design concerns for interactive hats. Through this lens, we detail the design and fabrication of three hat prototypes capable of sensing touch, head movements, and gestures, and including ambient displays of several types. Finally, we report an evaluation of our hat prototype and insights to inform the design of future hat technologies.

Millions of images are shared through social media every day. Yet, we know little about how the activities and preferences of users are dependent on the content of these images. In this paper, we seek to understand viewers engagement with photos. We design a quantitative study to expand previous research on in-app visual effects (also known as filters) through the examination of visual content identified through computer vision. This study is based on analysis of 4.9M Flickr images and is organized around three important engagement factors, likes, comments and favorites. We find that filtered photos are not equally engaging across different categories of content. Photos of food and people attract more engagement when filters are used, while photos of natural scenes and photos taken at night are more engaging when left unfiltered. In addition to contributing to the research around social media engagement and photography practices, our findings offer several design implications for mobile photo sharing platforms.

In multi-participant postings, as in online chat conversations, several conversations or topic threads may take place concurrently. This leads to difficulties for readers reviewing the postings in not only following discussions but also in quickly identifying their essence. A two-step process, disentanglement of interleaved posts followed by summarization of each thread, addresses the issue, but disentanglement errors are propagated to the summarization step, degrading the overall performance. To address this, we propose an end-to-end trainable encoder-decoder network for summarizing interleaved posts. The interleaved posts are encoded hierarchically, i.e., word-to-word (words in a post) followed by post-to-post (posts in a channel). The decoder also generates summaries hierarchically, thread-to-thread (generate thread representations) followed by word-to-word (i.e., generate summary words). Additionally, we propose a hierarchical attention mechanism for interleaved text. Overall, our end-to-end trainable hierarchical framework enhances performance over a sequence to sequence framework by 8\% on a synthetic interleaved texts dataset.

Advancements in 360° cameras have increased their related livestreams. In the case of video conferencing, 360° cameras provide almost unrestricted visibility into a conference room for a remote viewer without the need for an articulating camera. However, local participants are left wondering if someone is connected and where remote participants might be looking. To address this, we fabricated a prototype device that shows the gaze and presence of remote 360° viewers using a ring of LEDs that match the remote viewports. We discuss the long term use of one of the prototypes in a lecture hall and present future directions for visualizing gaze presence in 360° video streams.

Livestreaming and video calls have grown in popularity due to the increased connectivity and advancements in mobile devices. Our interactions with these cameras are limited as the cameras are either fixed or manually remote controlled. Here we present a Wizard-of-Oz elicitation study to inform the design of interactions with smart 360\textdegree\ cameras or robotic mobile desk cameras for use in video-conferences and live-streaming situations. There was an overall preference for devices that can minimize distraction as well as preferences for devices that can show they demonstrate an understanding of video-meeting context. We find participants dynamically grow with regards to the complexity of interactions which illustrate the need for deeper event semantics within the Camera AI. Finally, we detail interaction techniques and design insights to inform the next generation of personal video cameras for streaming and collaboration.

Reliable location estimation has been a key enabler of many applications in the UbiComp space. Much progress has been made on the development of accurate of indoor location systems, which form the foundation of many interesting applications, particularly in consumer scenarios. However, many location-based applications in enterprise settings also require addressing another facet of reliability: assurance. Without having strong guarantees of a location estimate’s legitimacy, stakeholders must explicitly balance the advantages offered with the risks of falsification. In this space, there are two key threats: replay attacks, where signal and sensor information is collected in one location and replayed in another to falsify a location estimation later in time; and wormhole attacks, where signal and sensor information is forwarded to a remote location by a colluding device to falsify location estimation in real-time. In this work, we improve upon the state of the art in wormhole-resistant location estimation techniques. Specifically, we present the Location Anchor, which leverages a combination of technical solutions and social contracts to provide high-assurance proofs of device location that are resistant to wormhole attacks. Unlike existing work, the Location Anchor has minimal hardware costs, supports a rich tapestry of applications, and is compatible with commodity smartphone and tablet platforms. We show that the Location Anchor can extend existing replay-resistant location systems into wormhole-resistant location systems, even in the face of very aggressive attacker assumptions. We describe the protocols underlying the Location Anchor, as well as report on the efficacy of a prototype implementation.

We describe a corpus analysis method to extract terminology from a collection of technical specifications book in the field of construction. Using statistics and word n-grams analyzes, we extract the terminology of the domain and then perform pruning steps with linguistic patterns and internet queries to improve the quality of the final terminology. In this paper we specifically focus on the improvements got by applying Internet queries and patterns. These improvements are evaluated by using a manual evaluation carried out by 6 experts in the field in the case of technical specification books.

Work breaks -- both physical and digital -- play an important role in productivity and workplace wellbeing. Yet, the growing availability of digital distractions from online content can turn breaks into prolonged "cyberloafing". In this paper, we present UpTime, a system that aims to support workers' transitions from breaks back to work--moments susceptible to digital distractions. Combining a browser extension and chatbot, users interact with UpTime through proactive and reactive chat prompts. By sensing transitions from inactivity, UpTime helps workers avoid distractions by automatically blocking distracting websites temporarily, while still giving them control to take necessary digital breaks. We report findings from a 3-week comparative field study with 15 workers. Our results show that automatic, temporary blocking at transition points can significantly reduce digital distractions and stress without sacrificing workers' sense of control. Our findings, however, also emphasize that overloading users' existing communication channels for chatbot interaction should be done thoughtfully.

Recent advances on the Internet of Things (IoT) lead to an explosion of physical objects being connected to the Internet. These objects sense, compute, interpret what is occurring within themselves and the world, and preferably interact with users. In this work, we present a visible light-enabled finger tracking technique allowing users to perform freestyle multi-touch gestures on everyday object’s surface. By projecting encoded patterns onto an object’s surface (e.g. paper, display, or table) through a projector, and localizing the user’s fingers with light sensors, the proposed system offers users a richer interactive space than the device’s existing interfaces. More importantly, results from our experiments indicate that this system can localize ten fingers simultaneously with an accuracy of 1.7 millimeters and an refresh rate of 84 Hz with only 31 milliseconds delay on WiFi or 23 milliseconds delay on serial communication, easily supporting multi-finger gesture interaction on everyday ob-jects. We also develop two example applications to demonstrate possible scenarios. Finally, we conduct a pre-liminary exploration of 3D depth inference using the same setup and achieve 2.43 cm depth estimation accuracy.

We present an approach to detect speech impairments from video of people with aphasia, a neurological condition that affects the ability to comprehend and produce speech. To counter inherent privacy issues, we propose a cross-media approach using only visual facial features to detect speech properties without listening to the audio content of speech. Our method uses facial landmark detections to measure facial motion over time. We show how to detect speech and pause instances based on temporal mouth shape analysis and identify repeating mouth patterns using a dynamic warping mechanism. We relate our developed features for pause frequency, mouth pattern repetitions, and pattern variety to actual symptoms of people with aphasia in the AphasiaBank dataset. Our evaluation shows that our developed features are able to reliably differentiate dysfluent speech production of people with aphasia from those without aphasia with an accuracy of 0.86. A combination of these handcrafted features and further statistical measures on talking and repetition improves classification performance to an accuracy of 0.88.

Activity recognition is a core component of many intelligent and context-aware systems. We present a solution for discreetly and unobtrusively recognizing common work activities above a work surface without using cameras.We demonstrate our approach, which utilizes an RF-radar sensor mounted under the work surface, in three domains; recognizing work activities at a convenience-store counter, recognizing common office deskwork activities, and estimating the position of customers in a showroom environment. Our examples illustrate potential benefits for both post-hoc business analytics and for real-time applications. Our solution was able to classify seven clerk activities with 94.9% accuracy using data collected in a lab environment and able to recognize six common deskwork activities collected in real offices with 95.3% accuracy. Using two sensors simultaneously, we demonstrate coarse position estimation around a large surface with 95.4% accuracy. We show that using multiple projections of RF signal leads to improved recognition accuracy. Finally, we show how smartwatches worn by users can be used to attribute an activity, recognized with the RF sensor, to a particular user in multi-user scenarios. We believe our solution can mitigate some of users’ privacy concerns associated with cameras and is useful for a wide range of intelligent systems.

Interaction in Augmented Reality or Mixed Reality environments is generally classified into two modalities: linear (relative to object) or non-linear (relative to camera). Switching between these modes can be arduous in cases where someone's interaction with the device is limited or restricted as is often the case in medical or industrial applications where one's hands might be sterile or soiled. To solve this, we present Sound-to-Experience where the modality can be effectively toggled by a noise or sound which is detected using a modern Artificial Intelligence deep-network classifier.

The analysis of bipartite networks is critical in many application domains, such as studying gene expression in bio-informatics. One important task is missing link prediction, which infers the exis- tence of new links based on currently observed ones. However, in practice, analysts need to utilize their domain knowledge based on the algorithm outputs in order to make sense of the results. We pro- pose a novel visual analysis framework, MissBi, which allows for examining and understanding missing links in bipartite networks. Some initial feedback from a management school professor has demonstrated the effectiveness of the tool.

Projector-camera systems can turn any surface such as tabletops and walls into an interactive display. A basic problem is to recognize the gesture actions on the projected UI widgets. Previous approaches using finger template matching or occlusion patterns have issues with environmental lighting conditions, artifacts and noise in the video images of a projection, and inaccuracies of depth cameras. In this work, we propose a new recognizer that employs a deep neural net with an RGB-D camera; specifically, we use a CNN (Convolutional Neural Network) with optical flow computed from the color and depth channels. We evaluated our method on a new dataset of RGB-D videos of 12 users interacting with buttons projected on a tabletop surface.

Searching collaboratively for places of interest is a common activity that frequently occurs on individual mobile phones, or on large tourist-information displays in public places such as visitor centers or train stations. We created a public display system for collaborative travel planning, as well as a mobile app that can augment the display. We tested them against third-party mobile apps in a simulated travel-search task to understand how the unique features of mobile phones and large displays might be leveraged together to improve collaborative travel planning experience.

We leverage a popularity measure in social media as a distant label for extractive summarization of online conversations. In social media, users can vote, share, or bookmark a post they prefer. The number of these actions is regarded as a measure of popularity. However, popularity is not solely determined by content of a post, e.g., a text or an image in a post, but is highly contaminated by its contexts, e.g., timing, and authority. We propose a disjunctive model, which computes the contribution of content and context separately. For evaluation, we build a dataset where the informativeness of a comment is annotated. We evaluate the results with ranking metrics, and show that our model outperforms the baseline model, which directly uses popularity as a measure of informativeness.

Prewriting is the process of generating and organizing ideas before drafting a document. Although often overlooked by novice writers and writing tool developers, prewriting is a critical process that improves the quality of a final document. To better understand current prewriting practices, we first conducted interviews with writing learners and experts. Based on the learners’ needs and experts’ recommendations, we then designed and developed InkPlanner, a novel pen and touch visualization tool that allows writers to utilize visual diagramming for ideation during prewriting. InkPlanner further allows writers to sort their ideas into a logical and sequential narrative by using a novel widget— NarrativeLine. Using a NarrativeLine, InkPlanner can automatically generate a document outline to guide later drafting exercises. Inkplanner is powered by machine-generated semantic and structural suggestions that are curated from various texts. To qualitatively review the tool and understand how writers use InkPlanner for prewriting, two writing experts were interviewed and a user study was conducted with university students. The results demonstrated that InkPlanner encouraged writers to generate more diverse ideas and also enabled them to think more strategically about how to organize their ideas for later drafting.

With the tremendous progress in sensing and IoT infrastructure, it is foreseeable that IoT systems will soon be available for commercial markets, such as in people's homes. In this paper, we present a deployment study using sensors attached to household objects to capture the resourcefulness of three individuals. The concept of resourcefulness highlights the ability of humans to repurpose objects spontaneously for a different use case than was initially intended. It is a crucial element for human health and wellbeing, which is of great interest for various aspects of HCI and design research. Traditionally, resourcefulness is captured through ethnographic practice. Ethnography can only provide sparse and often short duration observations of human experience, often relying on participants being aware of and remembering behaviours or thoughts they need to report on. Our hypothesis is that resourcefulness can also be captured through continuously monitoring objects being used in everyday life. We developed a system that can record object movement continuously and deployed them in homes of three elderly people for over two weeks. We explored the use of probabilistic topic models to analyze the collected data and identify common patterns.

Despite reflection being identified as a key component of behavior change, most existing tools do not explicitly design for it, carrying an implicit assumption that providing access to self-tracking data is enough to trigger reflection. In this work we design a system for reflection around physical activity. Through a set of workshops, we generated a corpus of 275 reflective questions. We then combine these questions into a set of 25 reflective mini-dialogues. We deliver our mini-dialogues through MMS. 33 active users of fitness trackers used our system in a 2-week field deployment. Results suggest that the mini-dialogues were successful in triggering reflection and that this reflection led to increases in motivation, empowerment, and adoption of new behaviors. Encouragingly, 16 participants elected to use the system for two additional weeks without compensation. We present implications for the design of technology-supported dialog system for reflection.