arXiv subjects

Jianan Pan

Publications and source records attributed to Jianan Pan.

6 recordsLinked to original sources

arXiv2026

PCOV-KWS: Multi-task Learning for Personalized Customizable Open Vocabulary Keyword Spotting

As advancements in technologies like Internet of Things (IoT), Automatic Speech Recognition (ASR), Speaker Verification (SV), and Text-to-Speech (TTS) lead to increased usage of intelligent voice assistants, the demand for privacy and personalization has escalated. In this paper, we introduce a multi-task learning framework for personalized, customizable open-vocabulary Keyword Spotting (PCOV-KWS). This framework employs a lightweight network to simultaneously perform Keyword Spotting (KWS) and SV to address personalized KWS requirements. We have integrated a training criterion distinct from softmax-based loss, transforming multi-class classification into multiple binary classifications, which eliminates inter-category competition, while an optimization strategy for multi-task loss weighting is employed during training. We evaluated our PCOV-KWS system in multiple datasets, demonstrating that it outperforms the baselines in evaluation results, while also requiring fewer parameters and lower computational resources.

eess.AS↗

arXiv2026

ProKWS: Personalized Keyword Spotting via Collaborative Learning of Phonemes and Prosody

Current keyword spotting systems primarily use phoneme-level matching to distinguish confusable words but ignore user-specific pronunciation traits like prosody (intonation, stress, rhythm). This paper presents ProKWS, a novel framework integrating fine-grained phoneme learning with personalized prosody modeling. We design a dual-stream encoder where one stream derives robust phonemic representations through contrastive learning, while the other extracts speaker-specific prosodic patterns. A collaborative fusion module dynamically combines phonemic and prosodic information, enhancing adaptability across acoustic environments. Experiments show ProKWS delivers highly competitive performance, comparable to state-of-the-art models on standard benchmarks and demonstrates strong robustness for personalized keywords with tone and intent variations.

eess.AS↗

arXiv2025

Holistic Information Theory of Spatial Remote Sensing Imaging

To address the non-optimal global design caused by the independent optimization of optical lenses, photodetectors, and computational processing subsystems in traditional remote sensing imaging system design, this paper proposes a holistic information theory for spatial remote sensing imaging. This theory integrates the optoelectronic imaging hardware front end and computational reconstruction back end into a unified framework. It establishes a complete spatial imaging chain information transfer model with the objective of obtaining the required effective information. The paper innovatively proposes a quantifiable Modulation Transfer Function (MTF)-Signal-to-Noise Ratio (SNR) product criterion. It demonstrates that the system information transmission ability is determined by the product of MTF and SNR, and that these parameters can compensate for each other to achieve equivalent information transfer. Validation through a high-resolution Earth observation system case shows that under consistent reconstruction mean square error conditions, increasing time delay integration stages reduces optical aperture size and significantly lowers primary mirror mass. Simulations and physical experiments further indicate that by increasing integration time, low-resolution optical systems achieve reconstructed fidelity comparable to high-resolution systems. This verifies that small-aperture optical systems can achieve equivalent imaging performance by enhancing SNR. This theory has been successfully applied in the design of the Jilin-1 satellite constellation, providing a new paradigm for low-cost high-resolution remote sensing systems.

physics.optics↗

arXiv2024

Estimating HIV Cross-sectional Incidence Using Recency Tests from a Non-representative Sample

Cross-sectional incidence estimation based on recency testing has become a widely used tool in HIV research. Recently, this method has gained prominence in HIV prevention trials to estimate the "placebo" incidence that participants might experience without preventive treatment. The application of this approach faces challenges due to non-representative sampling, as individuals aware of their HIV-positive status may be less likely to participate in screening for an HIV prevention trial. To address this, a recent phase 3 trial excluded individuals based on whether they have had a recent HIV test. To the best of our knowledge, the validity of this approach has yet to be studied. In our work, we investigate the performance of cross-sectional HIV incidence estimation when excluding individuals based on prior HIV tests in realistic trial settings. We develop a statistical framework that incorporates a testing-based criterion and possible non-representative sampling. We introduce a metric we call the effective mean duration of recent infection (MDRI) that mathematically quantifies bias in incidence estimation. We conduct an extensive simulation study to evaluate incidence estimator performance under various scenarios. Our findings reveal that when screening attendance is affected by knowledge of HIV status, incidence estimators become unreliable unless all individuals with recent HIV tests are excluded. Additionally, we identified a trade-off between bias and variability: excluding more individuals reduces bias from non-representative sampling but in many cases increases the variability of incidence estimates. These findings highlight the need for caution when applying testing-based criteria and emphasize the importance of refining incidence estimation methods to improve the design and evaluation of future HIV prevention trials.

stat.ME↗

arXiv2024

Spatio-temporal Joint Analysis of PM2.5 and Ozone in California with INLA

The substantial threat of concurrent air pollutants to public health is increasingly severe under climate change. To identify the common drivers and extent of spatio-temporal similarity of PM2.5 and ozone, this paper proposed a log Gaussian-Gumbel Bayesian hierarchical model allowing for sharing a SPDE-AR(1) spatio-temporal interaction structure. The proposed model outperforms in terms of estimation accuracy and prediction capacity for its increased parsimony and reduced uncertainty, especially for the shared ozone sub-model. Besides the consistently significant influence of temperature (positive), extreme drought (positive), fire burnt area (positive), and wind speed (negative) on both PM2.5 and ozone, surface pressure and GDP per capita (precipitation) demonstrate only positive associations with PM2.5 (ozone), while population density relates to neither. In addition, our results show the distinct spatio-temporal interactions and different seasonal patterns of PM2.5 and ozone, with peaks of PM2.5 and ozone in cold and hot seasons, respectively. Finally, with the aid of the excursion function, we see that the areas around the intersection of San Luis Obispo and Santa Barbara counties are likely to exceed the unhealthy ozone level for sensitive groups throughout the year. Our findings provide new insights for regional and seasonal strategies in the co-control of PM2.5 and ozone. Our methodology is expected to be utilized when interest lies in multiple interrelated processes in the fields of environment and epidemiology.

physics.ao-ph↗

arXiv2022

Adaptive Network Combination for Single-Image Reflection Removal: A Domain Generalization Perspective

Recently, multiple synthetic and real-world datasets have been built to facilitate the training of deep single image reflection removal (SIRR) models. Meanwhile, diverse testing sets are also provided with different types of reflection and scenes. However, the non-negligible domain gaps between training and testing sets make it difficult to learn deep models generalizing well to testing images. The diversity of reflections and scenes further makes it a mission impossible to learn a single model being effective to all testing sets and real-world reflections. In this paper, we tackle these issues by learning SIRR models from a domain generalization perspective. Particularly, for each source set, a specific SIRR model is trained to serve as a domain expert of relevant reflection types. For a given reflection-contaminated image, we present a reflection type-aware weighting (RTAW) module to predict expert-wise weights. RTAW can then be incorporated with adaptive network combination (AdaNEC) for handling different reflection types and scenes, i.e., generalizing to unknown domains. Two representative AdaNEC methods, i.e., output fusion (OF) and network interpolation (NI), are provided by considering both adaptation levels and efficiency. For images from one source set, we train RTAW to only predict expert-wise weights of other domain experts for improving generalization ability, while the weights of all experts are predicted and employed during testing. An in-domain expert (IDE) loss is presented for training RTAW. Extensive experiments show the appealing performance gain of our AdaNEC on different state-of-the-art SIRR networks. Source code and pre-trained models will available at https://github.com/csmliu/AdaNEC.

cs.CV↗