Xiaohui Liang

Youxiang Zhu, Nana Lin, Xiaohui Liang et al.

Using picture description speech for dementia detection has been studied for 30 years. Despite the long history, previous models focus on identifying the differences in speech patterns between healthy subjects and patients with dementia but do not utilize the picture information directly. In this paper, we propose the first dementia detection models that take both the picture and the description texts as inputs and incorporate knowledge from large pre-trained image-text alignment models. We observe the difference between dementia and healthy samples in terms of the text's relevance to the picture and the focused area of the picture. We thus consider such a difference could be used to enhance dementia detection accuracy. Specifically, we use the text's relevance to the picture to rank and filter the sentences of the samples. We also identified focused areas of the picture as topics and categorized the sentences according to the focused areas. We propose three advanced models that pre-processed the samples based on their relevance to the picture, sub-image, and focused areas. The evaluation results show that our advanced models, with knowledge of the picture and large image-text alignment models, achieve state-of-the-art performance with the best detection accuracy at 83.44%, which is higher than the text-only baseline model at 79.91%. Lastly, we visualize the sample and picture results to explain the advantages of our models.

Youxiang Zhu, Ruochen Li, Danqing Wang et al.

Long-context large language models (LLMs) are prone to be distracted by irrelevant contexts. The reason for distraction remains poorly understood. In this paper, we first identify the contextual heads, a special group of attention heads that control the overall attention of the LLM. Then, we demonstrate that distraction arises when contextual heads fail to allocate sufficient attention to relevant contexts and can be mitigated by increasing attention to these contexts. We further identify focus directions, located at the key and query activations of these heads, which enable them to allocate more attention to relevant contexts without explicitly specifying which context is relevant. We comprehensively evaluate the effect of focus direction on various long-context tasks and find out focus directions could help to mitigate the poor task alignment of the long-context LLMs. We believe our findings could promote further research on long-context LLM alignment.

Kristin Qi, Jiatong Shi, Caroline Summerour et al.

Mild Cognitive Impairment (MCI) is an early stage of Alzheimer's disease (AD), a form of neurodegenerative disorder. Early identification of MCI is crucial for delaying its progression through timely interventions. Existing research has demonstrated the feasibility of detecting MCI using speech collected from clinical interviews or digital devices. However, these approaches typically analyze data collected at limited time points, limiting their ability to identify cognitive changes over time. This paper presents a longitudinal study using voice assistant systems (VAS) to remotely collect seven-session speech data at three-month intervals across 18 months. We propose two methods to improve MCI detection and the prediction of cognitive changes. The first method incorporates historical data, while the second predicts cognitive changes at two time points. Our results indicate improvements when incorporating historical data: the average F1-score for MCI detection improves from 58.6% to 71.2% (by 12.6%) in the case of acoustic features and from 62.1% to 75.1% (by 13.0%) in the case of linguistic features. Additionally, the prediction of cognitive changes achieves an F1-score of 73.7% in the case of acoustic features. These results confirm the potential of VAS-based speech sessions for early detection of cognitive decline.

Kristin Qi, Youxiang Zhu, Xiaohui Liang

We present our approach to the PerAnsSumm Shared Task, which involves perspective span identification and perspective-aware summarization in community question-answering (CQA) threads. For span identification, we adopt ensemble learning that integrates three transformer models through averaging to exploit individual model strengths, achieving an 82.91% F1-score on test data. For summarization, we design a suite of Chain-of-Thought (CoT) prompting strategies that incorporate keyphrases and guide information to structure summary generation into manageable steps. To further enhance summary quality, we apply prompt optimization using the DSPy framework and supervised fine-tuning (SFT) on Llama-3 to adapt the model to domain-specific data. Experimental results on validation and test sets show that structured prompts with keyphrases and guidance improve summaries aligned with references, while the combination of prompt optimization and fine-tuning together yields significant improvement in both relevance and factuality evaluation metrics.

Nana Lin, Youxiang Zhu, Xiaohui Liang et al.

Mild cognitive impairment (MCI) is a major public health concern due to its high risk of progressing to dementia. This study investigates the potential of detecting MCI with spontaneous voice assistant (VA) commands from 35 older adults in a controlled setting. Specifically, a command-generation task is designed with pre-defined intents for participants to freely generate commands that are more associated with cognitive ability than read commands. We develop MCI classification and regression models with audio, textual, intent, and multimodal fusion features. We find the command-generation task outperforms the command-reading task with an average classification accuracy of 82%, achieved by leveraging multimodal fusion features. In addition, generated commands correlate more strongly with memory and attention subdomains than read commands. Our results confirm the effectiveness of the command-generation task and imply the promise of using longitudinal in-home commands for MCI detection.

Youxiang Zhu, Bang Tran, Xiaohui Liang et al.

Speech pause is an effective biomarker in dementia detection. Recent deep learning models have exploited speech pauses to achieve highly accurate dementia detection, but have not exploited the interpretability of speech pauses, i.e., what and how positions and lengths of speech pauses affect the result of dementia detection. In this paper, we will study the positions and lengths of dementia-sensitive pauses using adversarial learning approaches. Specifically, we first utilize an adversarial attack approach by adding the perturbation to the speech pauses of the testing samples, aiming to reduce the confidence levels of the detection model. Then, we apply an adversarial training approach to evaluate the impact of the perturbation in training samples on the detection model. We examine the interpretability from the perspectives of model accuracy, pause context, and pause length. We found that some pauses are more sensitive to dementia than other pauses from the model's perspective, e.g., speech pauses near to the verb "is". Increasing lengths of sensitive pauses or adding sensitive pauses leads the model inference to Alzheimer's Disease, while decreasing the lengths of sensitive pauses or deleting sensitive pauses leads to non-AD.

Yue Sun, Deqiang Xu, Zhuoming Huang et al.

We introduce LIDAUS Localization of IoT Device via Anchor UAV SLAM), an infrastructure-free, multi-stage SLAM system that utilizes an Unmanned Aerial Vehicle (UAV) to accurately localize IoT devices in a 3D indoor space where GPS signals are unavailable or weak, e.g., manufacturing factories, disaster sites, or smart buildings. The lack of GPS signals and infrastructure support makes most of the existing indoor localization systems not practical when localizing a large number of wireless IoT devices. In addition, safety concerns, access restriction, and simply the huge amount of IoT devices make it not practical for humans to manually localize and track IoT devices. To address these challenges, the UAV in our LIDAUS system conducts multi-stage 3D SLAM trips to localize devices based only on RSSIs, the most widely available measurement of the signals of almost all commodity IoT devices. The main novelties of the system include a weighted entropy-based clustering algorithm to select high quality RSSI observation locations, a 3D U-SLAM algorithm that is enhanced by deploying anchor beacons along the UAV's path, and the path planning based on Eulerian cycles on multi-layer grid graphs that model the space in exploring stage and Steiner tree paths in searching stages. Our simulations and experiments of Bluetooth IoT devices have demonstrated that the system can achieve high localization accuracy based only on RSSIs of commodity IoT devices.

Youxiang Zhu, Xiaohui Liang

In this paper, we exploit a Fully Convolutional Network (FCN) to analyze the audio data of spontaneous speech for dementia detection. A fully convolutional network accommodates speech samples with varying lengths, thus enabling us to analyze the speech sample without manual segmentation. Specifically, we first obtain the Mel Frequency Cepstral Coefficient (MFCC) feature map from each participant's audio data and convert the speech classification task on audio data to an image classification task on MFCC feature maps. Then, to solve the data insufficiency problem, we apply transfer learning by adopting a pre-trained backbone Convolutional Neural Network (CNN) model from the MobileNet architecture and the ImageNet dataset. We further build a convolutional layer to produce a heatmap using Otsu's method for visualization, enabling us to understand the impact of the time-series audio segments on the classification results. We demonstrate that our classification model achieves 66.7% over the testing dataset, 62.5% of the baseline model provided in the ADReSS challenge. Through the visualization technique, we can evaluate the impact of audio segments, such as filled pauses from the participants and repeated questions from the investigator, on the classification results.

Mohammad Hadian, Thamer Altuwaiyan, Xiaohui Liang et al.

In-home IoT devices play a major role in healthcare systems as smart personal assistants. They usually come with a voice-enabled feature to add an extra level of usability and convenience to elderly, disabled people, and patients. In this paper, we propose an efficient and privacy-preserving voice-based search scheme to enhance the efficiency and the privacy of in-home healthcare applications. We consider an application scenario where patients use the devices to record and upload their voice to servers and the caregivers search the interested voices of their patient's based on the voice content, mood, tone and background sound. Our scheme preserves the richness and privacy of voice data and enables accurate and efficient voice-based search, while in current systems that use speech recognition the richness and privacy of voice data are compromised. Specifically, our scheme achieves the privacy by employing a homomorphic encryption; only encrypted voice data is uploaded to the server who is unable to access the original voice data. In addition, our scheme enables the server to selectively and accurately respond to caregiver's queries on the voice data based on voice's feature similarity. We evaluate our scheme through real experiments and show that our scheme even with privacy preservation can successfully match similar voice data at an average accuracy of 80.8%.

Mohammad Hadian, Xiaohui Liang, Thamer Altuwaiyan et al.

Mobile healthcare system integrating wearable sensing and wireless communication technologies continuously monitors the users' health status. However, the mHealth system raises a severe privacy concern as the data it collects are private information, such as heart rate and blood pressure. In this paper, we propose an efficient and privacy-preserving mHealth data release approach for the statistic data with the objectives to preserve the unique patterns in the original data bins. The proposed approach adopts the bucket partition algorithm and the differential privacy algorithm for privacy preservation. A customized bucket partition algorithm is proposed to combine the database value bins into buckets according to certain conditions and parameters such that the patterns are preserved. The differential privacy algorithm is then applied to the buckets to prevent an attacker from being able to identify the small changes at the original data. We prove that the proposed approach achieves differential privacy. We also show the accuracy of the proposed approach through extensive simulations on real data. Real experiments show that our partitioning algorithm outperforms the state-of-the-art in preserving the patterns of the original data by a factor of 1.75.

Jiajun Zhang, Xiaohui Liang, Zhikun Zhang et al.

Wearable devices enable users to collect health data and share them with healthcare providers for improved health service. Since health data contain privacy-sensitive information, unprotected data release system may result in privacy leakage problem. Most of the existing work use differential privacy for private data release. However, they have limitations in healthcare scenarios because they do not consider the unique features of health data being collected from wearables, such as continuous real-time collection and pattern preservation. In this paper, we propose Re-DPoctor, a real-time health data releasing scheme with $w$-day differential privacy where the privacy of health data collected from any consecutive $w$ days is preserved. We improve utility by using a specially-designed partition algorithm to protect the health data patterns. Meanwhile, we improve privacy preservation by applying newly proposed adaptive sampling technique and budget allocation method. We prove that Re-DPoctor satisfies $w$-day differential privacy. Experiments on real health data demonstrate that our method achieves better utility with strong privacy guarantee than existing state-of-the-art methods.