David A. Shamma

Jingchao Fang, Nikos Arechiga, Keiichi Namaoshi et al.

The Wizard of Oz (WoZ) method is a widely adopted research approach where a human Wizard ``role-plays'' a not readily available technology and interacts with participants to elicit user behaviors and probe the design space. With the growing ability for modern large language models (LLMs) to role-play, one can apply LLMs as Wizards in WoZ experiments with better scalability and lower cost than the traditional approach. However, methodological guidance on responsibly applying LLMs in WoZ experiments and a systematic evaluation of LLMs' role-playing ability are lacking. Through two LLM-powered WoZ studies, we take the first step towards identifying an experiment lifecycle for researchers to safely integrate LLMs into WoZ experiments and interpret data generated from settings that involve Wizards role-played by LLMs. We also contribute a heuristic-based evaluation framework that allows the estimation of LLMs' role-playing ability in WoZ experiments and reveals LLMs' behavior patterns at scale.

Keiichi Namikoshi, Alex Filipowicz, David A. Shamma et al.

We consider the problem of aligning a large language model (LLM) to model the preferences of a human population. Modeling the beliefs, preferences, and behaviors of a specific population can be useful for a variety of different applications, such as conducting simulated focus groups for new products, conducting virtual surveys, and testing behavioral interventions, especially for interventions that are expensive, impractical, or unethical. Existing work has had mixed success using LLMs to accurately model human behavior in different contexts. We benchmark and evaluate two well-known fine-tuning approaches and evaluate the resulting populations on their ability to match the preferences of real human respondents on a survey of preferences for battery electric vehicles (BEVs). We evaluate our models against their ability to match population-wide statistics as well as their ability to match individual responses, and we investigate the role of temperature in controlling the trade-offs between these two. Additionally, we propose and evaluate a novel loss term to improve model performance on responses that require a numeric response.

Keiichi Namikoshi, David A. Shamma, Rumen Iliev et al.

Behavior change interventions are important to coordinate societal action across a wide array of important applications, including the adoption of electrified vehicles to reduce emissions. Prior work has demonstrated that interventions for behavior must be personalized, and that the intervention that is most effective on average across a large group can result in a backlash effect that strengthens opposition among some subgroups. Thus, it is important to target interventions to different audiences, and to present them in a natural, conversational style. In this context, an important emerging application domain for large language models (LLMs) is conversational interventions for behavior change. In this work, we leverage prior work on understanding values motivating the adoption of battery electric vehicles. We leverage new advances in LLMs, combined with a contextual bandit, to develop conversational interventions that are personalized to the values of each study participant. We use a contextual bandit algorithm to learn to target values based on the demographics of each participant. To train our bandit algorithm in an offline manner, we leverage LLMs to play the role of study participants. We benchmark the persuasive effectiveness of our bandit-enhanced LLM against an unaided LLM generating conversational interventions without demographic-targeted values.

Julie Williamson, Jie Li, Vinoba Vinayagamoorthy et al.

Virtual environments (VEs) can create collaborative and social spaces, which are increasingly important in the face of remote work and travel reduction. Recent advances, such as more open and widely available platforms, create new possibilities to observe and analyse interaction in VEs. Using a custom instrumented build of Mozilla Hubs to measure position and orientation, we conducted an academic workshop to facilitate a range of typical workshop activities. We analysed social interactions during a keynote, small group breakouts, and informal networking/hallway conversations. Our mixed-methods approach combined environment logging, observations, and semi-structured interviews. The results demonstrate how small and large spaces influenced group formation, shared attention, and personal space, where smaller rooms facilitated more cohesive groups while larger rooms made small group formation challenging but personal space more flexible. Beyond our findings, we show how the combination of data and insights can fuel collaborative spaces' design and deliver more effective virtual workshops.

David A. Shamma, Tony Dunnigan, Lyndon Kennedy

Photo applications offer tools for annotation via text and stickers. Ideophones, mimetic and onomatopoeic words, which are common in graphic novels, have yet to be explored for photo annotation use. We present a method for automatic ideophone recommendation and positioning of the text on photos. These annotations are accomplished by obtaining a list of ideophones with English definitions and applying a suite of visual object detectors to the image. Next, a semantic embedding maps the visual objects to the possible relevant ideophones. Our system stands in contrast to traditional computer vision-based annotation systems, which stop at recommending object and scene-level annotation, by providing annotations that are communicative, fun, and engaging. We test these annotations in Japanese and find they carry a strong preference and increase enjoyment and sharing likelihood when compared to unannotated and object-based annotated photos.

Yannis Kalantidis, Ayman Farahat, Lyndon Kennedy et al.

The quality of user experience online is affected by the relevance and placement of advertisements. We propose a new system for selecting and displaying visual advertisements in image search result sets. Our method compares the visual similarity of candidate ads to the image search results and selects the most visually similar ad to be displayed. The method further selects an appropriate location in the displayed image grid to minimize the perceptual visual differences between the ad and its neighbors. We conduct an experiment with about 900 users and find that our proposed method provides significant improvement in the users' overall satisfaction with the image search experience, without diminishing the users' ability to see the ad or recall the advertised brand.

Yannis Kalantidis, Lyndon Kennedy, Huy Nguyen et al.

We propose a novel hashing-based matching scheme, called Locally Optimized Hashing (LOH), based on a state-of-the-art quantization algorithm that can be used for efficient, large-scale search, recommendation, clustering, and deduplication. We show that matching with LOH only requires set intersections and summations to compute and so is easily implemented in generic distributed computing systems. We further show application of LOH to: a) large-scale search tasks where performance is on par with other state-of-the-art hashing approaches; b) large-scale recommendation where queries consisting of thousands of images can be used to generate accurate recommendations from collections of hundreds of millions of images; and c) efficient clustering with a graph-based algorithm that can be scaled to massive collections in a distributed environment or can be used for deduplication for small collections, like search results, performing better than traditional hashing approaches while only requiring a few milliseconds to run. In this paper we experiment on datasets of up to 100 million images, but in practice our system can scale to larger collections and can be used for other types of data that have a vector representation in a Euclidean space.

Ranjay Krishna, Yuke Zhu, Oliver Groth et al.

Despite progress in perceptual tasks such as image classification, computers still perform poorly on cognitive tasks such as image description and question answering. Cognition is core to tasks that involve not just recognizing, but reasoning about our visual world. However, models used to tackle the rich content in images for cognitive tasks are still being trained using the same datasets designed for perceptual tasks. To achieve success at cognitive tasks, models need to understand the interactions and relationships between objects in an image. When asked "What vehicle is the person riding?", computers will need to identify the objects in an image as well as the relationships riding(man, carriage) and pulling(horse, carriage) in order to answer correctly that "the person is riding a horse-drawn carriage". In this paper, we present the Visual Genome dataset to enable the modeling of such relationships. We collect dense annotations of objects, attributes, and relationships within each image to learn these models. Specifically, our dataset contains over 100K images where each image has an average of 21 objects, 18 attributes, and 18 pairwise relationships between objects. We canonicalize the objects, attributes, relationships, and noun phrases in region descriptions and questions answer pairs to WordNet synsets. Together, these annotations represent the densest and largest dataset of image descriptions, objects, attributes, relationships, and question answers.

Ranjay Krishna, Kenji Hata, Stephanie Chen et al.

Microtask crowdsourcing has enabled dataset advances in social science and machine learning, but existing crowdsourcing schemes are too expensive to scale up with the expanding volume of data. To scale and widen the applicability of crowdsourcing, we present a technique that produces extremely rapid judgments for binary and categorical labels. Rather than punishing all errors, which causes workers to proceed slowly and deliberately, our technique speeds up workers' judgments to the point where errors are acceptable and even expected. We demonstrate that it is possible to rectify these errors by randomizing task order and modeling response latency. We evaluate our technique on a breadth of common labeling tasks such as image verification, word similarity, sentiment analysis and topic classification. Where prior work typically achieves a 0.25x to 1x speedup over fixed majority vote, our approach often achieves an order of magnitude (10x) speedup.

Mohamed Kafsi, Henriette Cramer, Bart Thomee et al.

Geotagged data can be used to describe regions in the world and discover local themes. However, not all data produced within a region is necessarily specifically descriptive of that area. To surface the content that is characteristic for a region, we present the geographical hierarchy model (GHM), a probabilistic model based on the assumption that data observed in a region is a random mixture of content that pertains to different levels of a hierarchy. We apply the GHM to a dataset of 8 million Flickr photos in order to discriminate between content (i.e., tags) that specifically characterizes a region (e.g., neighborhood) and content that characterizes surrounding areas or more general themes. Knowledge of the discriminative and non-discriminative terms used throughout the hierarchy enables us to quantify the uniqueness of a given region and to compare similar but distant regions. Our evaluation demonstrates that our model improves upon traditional Naive Bayes classification by 47% and hierarchical TF-IDF by 27%. We further highlight the differences and commonalities with human reasoning about what is locally characteristic for a neighborhood, distilled from ten interviews and a survey that covered themes such as time, events, and prior regional knowledge

Bart Thomee, David A. Shamma, Gerald Friedland et al.

We present the Yahoo Flickr Creative Commons 100 Million Dataset (YFCC100M), the largest public multimedia collection that has ever been released. The dataset contains a total of 100 million media objects, of which approximately 99.2 million are photos and 0.8 million are videos, all of which carry a Creative Commons license. Each media object in the dataset is represented by several pieces of metadata, e.g. Flickr identifier, owner name, camera, title, tags, geo, media source. The collection provides a comprehensive snapshot of how photos and videos were taken, described, and shared over the years, from the inception of Flickr in 2004 until early 2014. In this article we explain the rationale behind its creation, as well as the implications the dataset has for science, research, engineering, and development. We further present several new challenges in multimedia research that can now be expanded upon with our dataset.