Study of Brain Connectivity by Multichannel EEG Quaternion Principal Component Analysis for Alzheimer Disease Classification

The early detection of Alzheimer's disease (AD) through widespread screening has emerged as a primary strategy to mitigate the significant global impact of AD. EEG measurements offer a promising solution for extensive AD detection. However, the intricate and nonlinear dynamics of multichannel EEG signals pose a considerable challenge for real-time AD diagnosis. This paper introduces a novel algorithm, which is based on Quaternion Principal Component Analysis (QPCA) of multichannel EEG signals, for AD classification. The algorithm extracts high dimensional correlations among different channels to generate features that are maximally representative with minimal information redundancy. This provides a multidimensional and precise measure of brain connectivity in disease assessment. Simulations have been conducted to evaluate the performance and to identify the most critical EEG channels or brain regions for AD classification. The results reveal a significant drop of connectivity measure in the alpha bands. The average AD classification accuracy for all 4-channel combinations reached 95%, while some particular permutations of channels achieved 100% accuracy rate. Furthermore, the temporal lobe emerges as one of the most important regions in AD classification given that the EEG signals are recorded during the presentation of an auditory stimulant. The selection of key parameters of the QPCA algorithm have been evaluated and some recommendations are proposed for further performance enhancement. This paper marks the first application of the QPCA algorithm for AD classification and brain connectivity analysis using multichannel EEG signals.

AcoustoBots: A swarm of robots for acoustophoretic multimodal interactions

Acoustophoresis has enabled novel interaction capabilities, such as levitation, volumetric displays, mid-air haptic feedback, and directional sound generation, to open new forms of multimodal interactions. However, its traditional implementation as a singular static unit limits its dynamic range and application versatility. This paper introduces AcoustoBots - a novel convergence of acoustophoresis with a movable and reconfigurable phased array of transducers for enhanced application versatility. We mount a phased array of transducers on a swarm of robots to harness the benefits of multiple mobile acoustophoretic units. This offers a more flexible and interactive platform that enables a swarm of acoustophoretic multimodal interactions. Our novel AcoustoBots design includes a hinge actuation system that controls the orientation of the mounted phased array of transducers to achieve high flexibility in a swarm of acoustophoretic multimodal interactions. In addition, we designed a BeadDispenserBot that can deliver particles to trapping locations, which automates the acoustic levitation interaction. These attributes allow AcoustoBots to independently work for a common cause and interchange between modalities, allowing for novel augmentations (e.g., a swarm of haptics, audio, and levitation) and bilateral interactions with users in an expanded interaction area. We detail our design considerations, challenges, and methodological approach to extend acoustophoretic central control in distributed settings. This work demonstrates a scalable acoustic control framework with two mobile robots, laying the groundwork for future deployment in larger robotic swarms. Finally, we characterize the performance of our AcoustoBots and explore the potential interactive scenarios they can enable.

Efficient Jailbreaking of Large Models by Freeze Training: Lower Layers Exhibit Greater Sensitivity to Harmful Content

With the widespread application of Large Language Models across various domains, their security issues have increasingly garnered significant attention from both academic and industrial communities. This study conducts sampling and normalization of the parameters of the LLM to generate visual representations and heatmaps of parameter distributions, revealing notable discrepancies in parameter distributions among certain layers within the hidden layers. Further analysis involves calculating statistical metrics for each layer, followed by the computation of a Comprehensive Sensitivity Score based on these metrics, which identifies the lower layers as being particularly sensitive to the generation of harmful content. Based on this finding, we employ a Freeze training strategy, selectively performing Supervised Fine-Tuning only on the lower layers. Experimental results demonstrate that this method significantly reduces training duration and GPU memory consumption while maintaining a high jailbreak success rate and a high harm score, outperforming the results achieved by applying the LoRA method for SFT across all layers. Additionally, the method has been successfully extended to other open-source large models, validating its generality and effectiveness across different model architectures. Furthermore, we compare our method with ohter jailbreak method, demonstrating the superior performance of our approach. By innovatively proposing a method to statistically analyze and compare large model parameters layer by layer, this study provides new insights into the interpretability of large models. These discoveries emphasize the necessity of continuous research and the implementation of adaptive security measures in the rapidly evolving field of LLMs to prevent potential jailbreak attack risks, thereby promoting the development of more robust and secure LLMs.

AbdomenAtlas: A Large-Scale, Detailed-Annotated, & Multi-Center Dataset for Efficient Transfer Learning and Open Algorithmic Benchmarking

We introduce the largest abdominal CT dataset (termed AbdomenAtlas) of 20,460 three-dimensional CT volumes sourced from 112 hospitals across diverse populations, geographies, and facilities. AbdomenAtlas provides 673K high-quality masks of anatomical structures in the abdominal region annotated by a team of 10 radiologists with the help of AI algorithms. We start by having expert radiologists manually annotate 22 anatomical structures in 5,246 CT volumes. Following this, a semi-automatic annotation procedure is performed for the remaining CT volumes, where radiologists revise the annotations predicted by AI, and in turn, AI improves its predictions by learning from revised annotations. Such a large-scale, detailed-annotated, and multi-center dataset is needed for two reasons. Firstly, AbdomenAtlas provides important resources for AI development at scale, branded as large pre-trained models, which can alleviate the annotation workload of expert radiologists to transfer to broader clinical applications. Secondly, AbdomenAtlas establishes a large-scale benchmark for evaluating AI algorithms -- the more data we use to test the algorithms, the better we can guarantee reliable performance in complex clinical scenarios. An ISBI & MICCAI challenge named BodyMaps: Towards 3D Atlas of Human Body was launched using a subset of our AbdomenAtlas, aiming to stimulate AI innovation and to benchmark segmentation accuracy, inference efficiency, and domain generalizability. We hope our AbdomenAtlas can set the stage for larger-scale clinical trials and offer exceptional opportunities to practitioners in the medical imaging community. Codes, models, and datasets are available at https://www.zongweiz.com/dataset

Generalized Mixture Model for Extreme Events Forecasting in Time Series Data

Time Series Forecasting (TSF) is a widely researched topic with broad applications in weather forecasting, traffic control, and stock price prediction. Extreme values in time series often significantly impact human and natural systems, but predicting them is challenging due to their rare occurrence. Statistical methods based on Extreme Value Theory (EVT) provide a systematic approach to modeling the distribution of extremes, particularly the Generalized Pareto (GP) distribution for modeling the distribution of exceedances beyond a threshold. To overcome the subpar performance of deep learning in dealing with heavy-tailed data, we propose a novel framework to enhance the focus on extreme events. Specifically, we propose a Deep Extreme Mixture Model with Autoencoder (DEMMA) for time series prediction. The model comprises two main modules: 1) a generalized mixture distribution based on the Hurdle model and a reparameterized GP distribution form independent of the extreme threshold, 2) an Autoencoder-based LSTM feature extractor and a quantile prediction module with a temporal attention mechanism. We demonstrate the effectiveness of our approach on multiple real-world rainfall datasets.