Yunqi Guo郭运奇

Earables, such as True Wireless Stereo earphones and VR/AR headsets, are increasingly popular, yet their compact design poses challenges for robust voice-related applications like telecommunication and voice assistant interactions in noisy environments. Existing speech enhancement systems, reliant solely on omnidirectional microphones, struggle with ambient noise like competing speakers. To address these issues, we propose VibOmni, a lightweight, end-to-end multi-modal speech enhancement system for earables that leverages bone-conducted vibrations captured by widely available Inertial Measurement Units (IMUs). VibOmni integrates a two-branch encoder-decoder deep neural network to fuse audio and vibration features. To overcome the scarcity of paired audio-vibration datasets, we introduce a novel data augmentation technique that models Bone Conduction Functions (BCFs) from limited …

Rural China faces rapid aging as out-migration weakens family support and formal services remain inaccessible. Community-based mutual aid emerges as a critical alternative, yet transaction-oriented models consistently fail. Through fieldwork in two contrasting villages involving 38 interviews and a participatory workshop, we investigate why quantifying care erodes the relational logic of rural mutual aid. We identify three dilemmas: weakening connections where depopulation makes needs invisible, suppressed help-seeking driven by dignity concerns and fear of unrepayable debt, and erosion of motivation as helping behaviors become socially invisible. Findings reveal that transaction-oriented platforms commodify care, creating “cold” exchanges that clash with “warm” relational norms. We propose a mediated reciprocity framework with three concepts: Relational Routing, Voice Mediation, and Symbolic …

In crowded social settings like conferences, background noise, overlapping voices, and lively interactions often lead to "cocktail party deafness," hindering clear conversation. While modern earphones are a promising platform for speech enhancement, existing solutions are limited: they either operate on a single device, ignoring the multi-party nature of conversation, or rely on impractical assumptions like fixed conversation areas and pre-recorded audio. We present CoHear, a collaborative system that leverages a network of earphones to holistically model and enhance speech at the conversation level. CoHear bridges acoustic sensor networks with deep learning for target speech extraction through two key contributions: 1) a novel, conversation-driven network that dynamically forms groups based on user interaction, using verbal and non-verbal cues (primarily head orientation) for robust, infrastructure-free coordination; and 2) a bandwidth-efficient, robust target speech extraction model that effectively utilizes peer-relayed audio as conditioning signals, even under network constraints. CoHear is evaluated in both real-world experiments and simulations. Results show that the conversation network obtains more than 90% accuracy in group formation, improves speech quality by up to 8.8 dB over state-of-the-art baselines, and demonstrates real-time performance on a mobile device. In a user study with 20 participants, CoHear achieved higher scores than baseline with good usability.

Sign-to-911 offers a compact mobile system solution to fast and runtime American Sign Language (ASL) and English translations. It is designated as 911 call services for ASL users with hearing disabilities upon emergencies. It enables bidirectional translations of ASL-to-English and English-to-ASL. The signer wears the AR glasses, runs Sign-to-911 on his/her smartphone and glasses, and interacts with a 911 operator. The design of Sign-to-911 departs from the popular deep learning based solution paradigm, and adopts simpler traditional AI/machine learning (ML) models. The key is to exploit ASL linguistic features to simplify the model structures and improve accuracy and speed. It further leverages recent component solutions from graphics, vision, natural language processing, and AI/ML. Our evaluation with six ASL signers and 911 call records has confirmed its viability.