Existing recommendation systems either rely on user interaction logs, such as online shopping history for shopping recommendations, or focus on text signals. However, item-based histories are not always accessible and generalizable for multimodal recommendation. We hypothesize that a user's visual history --- comprising images from daily life --- can offer rich, task-agnostic insights into their interests and preferences, and thus be leveraged for effective personalization. To this end, we propose VisualLens, a novel framework that leverages multimodal large language models (MLLMs) to enable personalization using task-agnostic visual history. VisualLens extracts, filters, and refines a spectrum user profile from the visual history to support personalized recommendation. We created two new benchmarks, Google-Review-V and Yelp-V, with task-agnostic visual histories, and show that VisualLens improves over state-of-the-art item-based multimodal recommendations by 5-10\% on Hit@3, and outperforms GPT-4o by 2-5\%. Further analysis shows that VisualLens is robust across varying history lengths and excels at adapting to both longer histories and unseen content categories.

Self-supervised learning (SSL) has shown impressive results in downstream classification tasks. However, there is limited work in understanding their failure modes and interpreting their learned representations. In this paper, we study the representation space of state-of-the-art self-supervised models including SimCLR, SwaV, MoCo, BYOL, DINO, SimSiam, VICReg and Barlow Twins. Without the use of class label information, we discover discriminative features that correspond to unique physical attributes in images, present mostly in correctly-classified representations. Using these features, we can compress the representation space by up to$40% without significantly affecting linear classification performance. We then propose Self-Supervised Representation Quality Score (or Q-Score), an unsupervised score that can reliably predict if a given sample is likely to be mis-classified during linear evaluation, achieving AUPRC of 91.45 on ImageNet-100 and 78.78 on ImageNet-1K. Q-Score can also be used as a regularization term on pre-trained encoders to remedy low-quality representations. Fine-tuning with Q-Score regularization can boost the linear probing accuracy of SSL models by up to 5.8% on ImageNet-100 and 3.7% on ImageNet-1K compared to their baselines. Finally, using gradient heatmaps and Salient ImageNet masks, we define a metric to quantify the interpretability of each representation. We show that discriminative features are strongly correlated to core attributes and, enhancing these features through Q-score regularization makes SSL representations more interpretable.

The quality variance in user-generated content is a major bottleneck to serving communities on online platforms. Current content ranking methods primarily evaluate text and non-textual content features of each user post in isolation. In this paper, we demonstrate the utility of considering the implicit and explicit relational aspects across user content to assess their quality. First, we develop a modular platform-agnostic framework to represent the contrastive (or competing) and similarity-based relational aspects of user-generated content via independently induced content graphs. Second, we develop two complementary graph convolutional operators that enable feature contrast for competing content and feature smoothing/sharing for similar content. Depending on the edge semantics of each content graph, we embed its nodes via one of the above two mechanisms. We also show that our contrastive operator creates discriminative magnification across the embeddings of competing posts. Third, we show a surprising result-applying classical boosting techniques to combine final-layer embeddings across the content graphs significantly outperforms the typical stacking, fusion, or neighborhood embedding aggregation methods in graph convolutional architectures. We exhaustively validate our method via accepted answer prediction over fifty diverse Stack-Exchange (https://stackexchange.com/) websites with consistent relative gains of over 5% accuracy over state-of-the-art neural, multi-relational and textual baselines.

Email continues to be an important form of communication as well as a way to manage tasks and archive personal information. As the volume of email grows, organizing and finding relevant email remains challenging. In this paper, we present a large-scale log analysis of the activities that people perform on email mes-sages (accessing external information via links or attachments, responding to messages, and organizing messages), their search behavior, and their organizational practices in a popular web email client.

This paper lays the ground work for assistive navigation using wearable sensors and social sensors to foster situational awareness for the blind. Our system acquires social media messages to gauge the relevant aspects of an event and to create alerts. We propose social semantics that captures the parameters required for querying and reasoning an event-of-interest, such as what, where, who, when, severity, and action from the Internet of things, using an event summarization algorithm. Our approach integrates wearable sensors in the physical world to estimate user location based on metric and landmark localization. Streaming data from the cyber world are employed to provide awareness by summarizing the events around the user based on the situation awareness factor. It is illustrated using disaster and socialization event scenarios. Discovered local events are fed back using sound localization so that the user can actively participate in a social event or get early warning of any hazardous events. A feasibility evaluation of our proposed algorithm included comparing the output of the algorithm to ground truth, a survey with sighted participants about the algorithm output, and a sound localization user interface study with blind-folded sighted participants. Thus, our framework supports the navigation problem for the blind by combining the advantages of our real-time localization technologies so that the user is being made aware of the world, a necessity for independent travel.