Deep Learning-based Framework Research Articles

DRLAR: A deep reinforcement learning-based adaptive routing framework for network-on-chips

Adaptive routing plays a pivotal role in the overall performance of Network-on-Chips (NoCs). However, with many-core architectures supporting complex and constantly changing traffic patterns for emerging applications, this aspect presents significant challenges. Our meticulous examination and analysis of existing heuristic adaptive routing algorithms revealed three key limiting factors: single network status metric, lack of system feedback awareness, and lack of ability to customize. Reinforcement Learning (RL) methods have demonstrated promising opportunities for exploring adaptive routing design. Deep reinforcement learning (DRL) techniques, in particular, enable efficient exploration in adaptive routing design spaces where heuristic strategies may be inadequate. This paper proposes a novel deep reinforcement learning framework for adaptive routing called DRLAR that is suitable for diversified traffic patterns and resolves multi-objective optimization simultaneously. DRLAR formulates routing as an agent, which makes routing decisions based on autonomous learning using multiple network state features and system-level feedback information. We conduct extensive experiments against state-of-the-art routing algorithms to evaluate our design. The results show that DRLAR reduces packet latency by an average of 33.3%, achieving a reduction of 41.6% and 10.5% on average under heavy synthetic traffic and the PARSEC 2.1 benchmark, respectively. We also perform a cost analysis to validate the potential implementation of DRLAR on NoCs with low computational, storage, and power.

CSST Dense Star Field Preparation: A Framework for Astrometry and Photometry for Dense Star Field Images Obtained by the China Space Station Telescope (CSST)

Abstract The Chinese Space Station Telescope (CSST) is a telescope with 2-meter diameter, obtaining images with high quality through wide-field observations. In its first observation cycle, the CSST will scan portions of the galactic centre with 7 different bands across different epochs to capture time-domain observation data. These data have significant potential for the study of properties of stars and exoplanets. However, the density of stars in the galactic centre is high, and it is a well-known challenge to perform astrometry and photometry in such a dense star field. This paper presents a deep learning-based framework designed to process dense star field images obtained by the CSST, which includes photometry, astrometry, and classifications of targets according to their light curve periods. With simulated CSST observation data, we demonstrate that this deep learning framework achieves photometry accuracy of 0.23% and astrometry accuracy of 0.03 pixel for stars with moderate brightness mag=24 in i band, surpassing results obtained by traditional methods. Additionally, the
deep learning based light curve classification algorithm could pick up celestial targets whose magnitude variations are 1.7 times larger than magnitude variations brought by Poisson Photon Noise. We anticipate that our framework could be effectively used
to process dense star field images obtained by the CSST.

Artificial intelligence enhances whole-slide interpretation ofPD-L1 CPS in triple-negative breast cancer: Amulti-institutional ring study.

Evaluation of the programmed cell death ligand-1 (PD-L1) combined positive score (CPS) is vital to predict the efficacy of the immunotherapy in triple-negative breast cancer (TNBC), but pathologists show substantial variability in the consistency and accuracy of the interpretation. It is of great importance to establish an objective and effective method which is highly repeatable. We proposed a model in a deep learning-based framework, which at the patch level incorporated cell analysis and tissue region analysis, followed by the whole-slide level fusion of patch results. Three rounds of ring studies (RSs) were conducted. Twenty-one pathologists of different levels from four institutions evaluated the PD-L1 CPS in TNBC specimens as continuous scores by visual assessment and our artificial intelligence (AI)-assisted method. In the visual assessment, the interpretation results of PD-L1 (Dako 22C3) CPS by different levels of pathologists have significant differences and showed weak consistency. Using AI-assisted interpretation, there were no significant differences between all pathologists (P = 0.43), and the intraclass correlation coefficient (ICC) value was increased from 0.618 [95% confidence interval (CI) = 0.524-0.719] to 0.931 (95% CI = 0.902-0.955). The accuracy of interpretation result is further improved to 0.919 (95% CI = 0.886-0.947). Acceptance of AI results by junior pathologists was the highest among all levels, and 80% of the AI results were accepted overall. With the help of the AI-assisted diagnostic method, different levels of pathologists achieved excellent consistency and repeatability in the interpretation of PD-L1 (Dako 22C3) CPS. Our AI-assisted diagnostic approach was proved to strengthen the consistency and repeatability in clinical practice.

Ensembles of multiple models for soil moisture retrieval from remote sensing data over agricultural areas: A deep learning-based framework

In agricultural areas, most surface soil moisture (SM) retrieval models are unstable in terms of their accuracy and performance during crop growth. As a result, there is no consensus on which model performs optimally during the agricultural season. This is because of the uncertainties associated with model physics, initial conditions, model inversion processes, input data, vegetation attenuation and soil characteristics. To better deal with these practical concerns, we propose a simple, but robust SM retrieval method of using combination of multiple models based on deep learning and multi-model ensemble approach, called the DL-MME method, which makes use of the ‘collective intelligence’ and ‘wisdom of crowds’ concepts. The advantages of this method are: (1) robustness to model selection, and (2) robustness to model calibration during the growing season. In addition, this method is less dependent on one type of data across various agricultural areas compared to the single model approach. Firstly, the coupled water cloud model (WCM) and soil backscattering models (Oh model or advanced integral equation model (AIEM)) with different vegetation descriptors were calibrated and validated during the growing season in sugarcane and winter wheat fields for Sentinel-1 backscattering coefficients (VV and VH). SM was also retrieved by employing the trapezoid model (OPTRAM) with different parameters from Sentinel-2 images. To optimize SM retrieval computations, we used the outputs from optical and SAR models, auxiliary features, and reliable in situ SM measurements as inputs to a deep learning convolutional neural network (DL-CNN). For sugarcane and wheat fields in the early stages of crop growth, WCM models retrieved more accurate time-series SM than optical models. OPTRAM soil moisture retrievals showed greater accuracy in the late crop growing season. Time-series SM retrieval accuracy using DL-MME was higher than for the optical and semi-empirical SAR models. According to the results of the in situ validation for wheat (sugarcane) fields, the minimum MAE by an optimal combination of models was around 0.01 (0.02) m3m−3 (RMSE = 0.036 (0.074) m3m−3; R = 0.87 (0.71)). The findings demonstrate that our method is reliable and feasible for SM retrieval. Additionally, our method provides a way to select an optimal model for retrieving time-series SM during the crop growing season.

Fake User Detection Based on Multi-Model Joint Representation

The existing deep learning-based detection of fake information focuses on the transient detection of news itself. Compared to user category profile mining and detection, transient detection is prone to higher misjudgment rates due to the limitations of insufficient temporal information, posing new challenges to social public opinion monitoring tasks such as fake user detection. This paper proposes a multimodal aggregation portrait model (MAPM) based on multi-model joint representation for social media platforms. It constructs a deep learning-based multimodal fake user detection framework by analyzing user behavior datasets within a time retrospective window. It integrates a pre-trained Domain Large Model to represent user behavior data across multiple modalities, thereby constructing a high-generalization implicit behavior feature spectrum for users. In response to the tendency of existing fake user behavior mining to neglect time-series features, this study introduces an improved network called Sequence Interval Detection Net (SIDN) based on Sequence to Sequence (seq2seq) to characterize time interval sequence behaviors, achieving strong expressive capabilities for detecting fake behaviors within the time window. Ultimately, the amalgamation of latent behavioral features and explicit characteristics serves as the input for spectral clustering in detecting fraudulent users. The experimental results on Weibo real dataset demonstrate that the proposed model outperforms the detection utilizing explicit user features, with an improvement of 27.0% in detection accuracy.

PERCORE: A Deep Learning-Based Framework for Persian Spelling Correction with Phonetic Analysis

This research introduces a state-of-the-art Persian spelling correction system that seamlessly integrates deep learning techniques with phonetic analysis, significantly enhancing the accuracy and efficiency of natural language processing (NLP) for Persian. Utilizing a fine-tuned language representation model, our methodology effectively combines deep contextual analysis with phonetic insights, adeptly correcting both non-word and real-word spelling errors. This strategy proves particularly effective in tackling the unique complexities of Persian spelling, including its elaborate morphology and the challenge of homophony. A thorough evaluation on a wide-ranging dataset confirms our system’s superior performance compared to existing methods, with impressive F1-Scores of 0.890 for detecting real-word errors and 0.905 for correcting them. Additionally, the system demonstrates a strong capability in non-word error correction, achieving an F1-Score of 0.891. These results illustrate the significant benefits of incorporating phonetic insights into deep learning models for spelling correction. Our contributions not only advance Persian language processing by providing a versatile solution for a variety of NLP applications but also pave the way for future research in the field, emphasizing the critical role of phonetic analysis in developing effective spelling correction system.

Development of a deep learning-based group contribution framework for targeted design of ionic liquids

In this article, we present a novel deep learning-based group contribution framework for the targeted design of ionic liquids (ILs). This computational framework can expedite and improve the process of finding desirable molecular structures of IL via accurate property predictions in a data-driven manner. Our proposed framework consists of two essential steps: establishing a correlation between IL viscosity and CO2 solubility by merging two deep learning models (DNN-GC and ANN-GC) and utilizing this correlation to identify the optimal IL structure with maximal CO2 absorption capacity. Our model achieves high accuracy with R2 values of 95%, 94.2%, and 96.4% for DNN-GC, ANN-GC, and DNN-ANN-GC, respectively. Correlation results align with the experimental data, affirming the applicability of our framework. Finally, the algorithm is employed in a CO2 capture case study to generate and select the best-performing novel ILs, which exhibit behavior consistent with established ILs in the literature.

Development of an Automated Morphometric Approach to Assess Vascular Outcomes following Exposure to Environmental Chemicals in Zebrafish.

Disruptions in vascular formation attributable to chemical insults is a pivotal risk factor or potential etiology of developmental defects and various disease settings. Among the thousands of chemicals threatening human health, the highly concerning groups prevalent in the environment and detected in biological monitoring in the general population ought to be prioritized because of their high exposure risks. However, the impacts of a large number of environmental chemicals on vasculature are far from understood. The angioarchitecture complexity and technical limitations make it challenging to analyze the entire vasculature efficiently and identify subtle changes through a high-throughput in vivo assay. We aimed to develop an automated morphometric approach for the vascular profile and assess the vascular morphology of health-concerning environmental chemicals. High-resolution images of the entire vasculature in Tg(fli1a:eGFP) zebrafish were collected using a high-content imaging platform. We established a deep learning-based quantitative framework, ECA-ResXUnet, combined with MATLAB to segment the vascular networks and extract features. Vessel scores based on the rates of morphological changes were calculated to rank vascular toxicity. Potential biomarkers were identified by vessel-endothelium-gene-disease integrative analysis. Whole-trunk blood vessels and the cerebral vasculature in larvae exposed to 150 representative chemicals were automatically segmented as comparable to human-level accuracy, with sensitivity and specificity of 95.56% and 95.81%, respectively. Chemical treatments led to heterogeneous vascular patterns manifested by 31 architecture indexes, and the common cardinal vein (CCV) was the most affected vessel. The antipsychotic medicine haloperidol, flame retardant 2,2-bis(chloromethyl)trimethylenebis[bis(2-chloroethyl) phosphate], and tert-butylphenyl diphenyl phosphate ranked as the top three in vessel scores. Pesticides accounted for the largest group, with a vessel score of , characterized by a remarkable inhibition of subintestinal venous plexus and delayed development of CCV. Multiple-concentration evaluation of nine per- and polyfluoroalkyl substances (PFAS) indicated a low-concentration effect on vascular impairment and a positive association between carbon chain length and benchmark concentration. Target vessel-directed single-cell RNA sequencing of cells from larvae treated with , perfluorohexanesulfonic acid, or benzylbutyl phthalate, along with vessel-endothelium-gene-disease integrative analysis, uncovered potential associations with vascular disorders and identified biomarker candidates. This study provides a novel paradigm for phenotype-driven screenings of vascular-disrupting chemicals by converging morphological and transcriptomic profiles at a high-resolution level, serving as a powerful tool for large-scale toxicity tests. Our approach and the high-quality morphometric data facilitate the precise evaluation of vascular effects caused by environmental chemicals. https://doi.org/10.1289/EHP13214.

DeepFace: Deep learning-based framework to contextualize orofacial cleft-related variants during human embryonic craniofacial development

Orofacial clefts (OFC) are among the most common human congenital birth defects. Previous multiethnic studies have identified dozens of associated loci for both cleft lip with or without cleft palate (CL/P) and cleft palate alone (CP). Although several nearby genes have been highlighted, the ‘casual’ variants are largely unknown. Here, we developed DeepFace, a convolutional neural network model, to assess the functional impact of variants by SNP activity difference (SAD) scores. The DeepFace model is trained with 204 epigenomic assays from crucial human embryonic craniofacial developmental stages of post-conception week (pcw) 4 to pcw 10. The Pearson Correlation Coefficient between the predicted and actual values for 12 epigenetic features achieved a median range of 0.50 to 0.83. Specifically, our model revealed that SNPs significantly associated with OFC tended to exhibit higher SAD scores across various variant categories compared to less-related groups, indicating a context-specific impact of OFC-related SNPs. Notably, we identified six SNPs with a significant linear relationship to SAD scores throughout developmental progression, suggesting that these SNPs could play a temporal regulatory role. Furthermore, our cell-type specificity analysis pinpointed the trophoblast cell as having the highest enrichment of risk signals associated with OFC. Overall, DeepFace can harness distal regulatory signals from extensive epigenomic assays, offering new perspectives for prioritizing OFC variants using contextualized functional genomic features. We expect DeepFace to be instrumental in accessing and predicting the regulatory roles of variants associated with OFC, and the model that can be extended to study other complex diseases or traits.

Fake Currency Detection Using Deep Learning

Fake currency is the money produced without the approval of the government, creation of it is considered as a great offence. The progression of shading printing innovation has expanded the rate of Fake currency copying notes on a large scale. Albeit electronic monetary exchanges are turning out to be more popular and the utilization of paper cash has been diminishing as of late, banknotes still remain in distribution attributable to their dependability and straight forwardness in use. Few years ago, the printing should be possible in a printing-houses, yet presently anybody can print a money paper with most extreme exactness utilizing a straightforward laser printer. As an outcome, the issue of duplicate currency rather than the authentic ones has been increases generally. India had reviled the problems like defilement and dark cash and fake of money notes is likewise a big issue to it. To handle this problem, a deep learning-based framework is proposed to identify the fake Indian currency. The MATLAB tool has been used to identify the fake currency. The outcome will classify whether the Indian currency note is Real or Fake

A physics-embedded deep-learning framework for efficient multi-fidelity modeling applied to guided wave based structural health monitoring

Health monitoring of structures using ultrasonic guided waves is an evolving technology with potential applications in monitoring pipelines, civil bridges, and aircraft components. However, the sensitivity of guided waves to external parameters affects the reliability of monitoring systems based on them. These influencing factors and experimental related factors cannot be perfectly modeled, which give rise to the discrepancy between numerical simulations and experimental measurements. Therefore, it is important to address this inevitable discrepancy and generate close-to-experiment simulations. In this work, we present a deep learning-based Digital Twin framework containing multi-fidelity modeling to reduce the discrepancy between measurements and simulations and a deep generative model to generate close-to-experiment guided wave responses by harnessing the vital characteristics of the two sources. These realistic simulations (close to experiment) can then be used in assessing the reliability of health monitoring system by generating probability of detection curves. Furthermore, they can also be used for augmenting the training data for a machine learning algorithm. We use a measurement dataset corresponding to crack propagation and simulations to validate the proposed framework. The results show that the discrepancy is indeed reduced to a great extent, furthermore, we also show that this framework enables the computation of probability of detection from close-to-experiment data as a direct consequence of rapid generation of realistic simulations.

Associations of Intrapancreatic Fat Deposition With Incident Diseases of the Exocrine and Endocrine Pancreas: A UK Biobank Prospective Cohort Study.

To investigate whether increased intrapancreatic fat deposition (IPFD) heightens the risk of diseases of the exocrine and endocrine pancreas. A prospective cohort study was conducted using data from the UK Biobank. IPFD was quantified using MRI and a deep learning-based framework called nnUNet. The prevalence of fatty change of the pancreas (FP) was determined using sex- and age-specific thresholds. Associations between IPFD and pancreatic diseases were assessed with multivariate Cox-proportional hazard model adjusted for age, sex, ethnicity, body mass index, smoking and drinking status, central obesity, hypertension, dyslipidemia, liver fat content, and spleen fat content. Of the 42,599 participants included in the analysis, the prevalence of FP was 17.86%. Elevated IPFD levels were associated with an increased risk of acute pancreatitis (hazard ratio [HR] per 1 quintile change 1.513, 95% confidence interval [CI] 1.179-1.941), pancreatic cancer (HR per 1 quintile change 1.365, 95% CI 1.058-1.762) and diabetes mellitus (HR per 1 quintile change 1.221, 95% CI 1.132-1.318). FP was also associated with a higher risk of acute pancreatitis (HR 3.982, 95% CI 2.192-7.234), pancreatic cancer (HR 1.976, 95% CI 1.054-3.704), and diabetes mellitus (HR 1.337, 95% CI 1.122-1.593, P = 0.001). FP is a common pancreatic disorder. Fat in the pancreas is an independent risk factor for diseases of both the exocrine pancreas and endocrine pancreas.

Predict lncRNA-drug associations based on graph neural network.

LncRNAs are an essential type of non-coding RNAs, which have been reported to be involved in various human pathological conditions. Increasing evidence suggests that drugs can regulate lncRNAs expression, which makes it possible to develop lncRNAs as therapeutic targets. Thus, developing in-silico methods to predict lncRNA-drug associations (LDAs) is a critical step for developing lncRNA-based therapies. In this study, we predict LDAs by using graph convolutional networks (GCN) and graph attention networks (GAT) based on lncRNA and drug similarity networks. Results show that our proposed method achieves good performance (average AUCs > 0.92) on five datasets. In addition, case studies and KEGG functional enrichment analysis further prove that the model can effectively identify novel LDAs. On the whole, this study provides a deep learning-based framework for predicting novel LDAs, which will accelerate the lncRNA-targeted drug development process.

A deep learning framework for identifying and segmenting three vessels in fetal heart ultrasound images

BackgroundCongenital heart disease (CHD) is one of the most common birth defects in the world. It is the leading cause of infant mortality, necessitating an early diagnosis for timely intervention. Prenatal screening using ultrasound is the primary method for CHD detection. However, its effectiveness is heavily reliant on the expertise of physicians, leading to subjective interpretations and potential underdiagnosis. Therefore, a method for automatic analysis of fetal cardiac ultrasound images is highly desired to assist an objective and effective CHD diagnosis.MethodIn this study, we propose a deep learning-based framework for the identification and segmentation of the three vessels—the pulmonary artery, aorta, and superior vena cava—in the ultrasound three vessel view (3VV) of the fetal heart. In the first stage of the framework, the object detection model Yolov5 is employed to identify the three vessels and localize the Region of Interest (ROI) within the original full-sized ultrasound images. Subsequently, a modified Deeplabv3 equipped with our novel AMFF (Attentional Multi-scale Feature Fusion) module is applied in the second stage to segment the three vessels within the cropped ROI images.ResultsWe evaluated our method with a dataset consisting of 511 fetal heart 3VV images. Compared to existing models, our framework exhibits superior performance in the segmentation of all the three vessels, demonstrating the Dice coefficients of 85.55%, 89.12%, and 77.54% for PA, Ao and SVC respectively.ConclusionsOur experimental results show that our proposed framework can automatically and accurately detect and segment the three vessels in fetal heart 3VV images. This method has the potential to assist sonographers in enhancing the precision of vessel assessment during fetal heart examinations.

A neural speech decoding framework leveraging deep learning and speech synthesis

Decoding human speech from neural signals is essential for brain–computer interface (BCI) technologies that aim to restore speech in populations with neurological deficits. However, it remains a highly challenging task, compounded by the scarce availability of neural signals with corresponding speech, data complexity and high dimensionality. Here we present a novel deep learning-based neural speech decoding framework that includes an ECoG decoder that translates electrocorticographic (ECoG) signals from the cortex into interpretable speech parameters and a novel differentiable speech synthesizer that maps speech parameters to spectrograms. We have developed a companion speech-to-speech auto-encoder consisting of a speech encoder and the same speech synthesizer to generate reference speech parameters to facilitate the ECoG decoder training. This framework generates natural-sounding speech and is highly reproducible across a cohort of 48 participants. Our experimental results show that our models can decode speech with high correlation, even when limited to only causal operations, which is necessary for adoption by real-time neural prostheses. Finally, we successfully decode speech in participants with either left or right hemisphere coverage, which could lead to speech prostheses in patients with deficits resulting from left hemisphere damage.

Design and Experimental Validation of Deep Reinforcement Learning-Based Fast Trajectory Planning and Control for Mobile Robot in Unknown Environment.

This article is concerned with the problem of planning optimal maneuver trajectories and guiding the mobile robot toward target positions in uncertain environments for exploration purposes. A hierarchical deep learning-based control framework is proposed which consists of an upper level motion planning layer and a lower level waypoint tracking layer. In the motion planning phase, a recurrent deep neural network (RDNN)-based algorithm is adopted to predict the optimal maneuver profiles for the mobile robot. This approach is built upon a recently proposed idea of using deep neural networks (DNNs) to approximate the optimal motion trajectories, which has been validated that a fast approximation performance can be achieved. To further enhance the network prediction performance, a recurrent network model capable of fully exploiting the inherent relationship between preoptimized system state and control pairs is advocated. In the lower level, a deep reinforcement learning (DRL)-based collision-free control algorithm is established to achieve the waypoint tracking task in an uncertain environment (e.g., the existence of unexpected obstacles). Since this approach allows the control policy to directly learn from human demonstration data, the time required by the training process can be significantly reduced. Moreover, a noisy prioritized experience replay (PER) algorithm is proposed to improve the exploring rate of control policy. The effectiveness of applying the proposed deep learning-based control is validated by executing a number of simulation and experimental case studies. The simulation result shows that the proposed DRL method outperforms the vanilla PER algorithm in terms of training speed. Experimental videos are also uploaded, and the corresponding results confirm that the proposed strategy is able to fulfill the autonomous exploration mission with improved motion planning performance, enhanced collision avoidance ability, and less training time.

Deep Multimodal Fusion of Data with Heterogeneous Dimensionality via Projective Networks.

The use of multimodal imaging has led to significant improvements in the diagnosis and treatment of many diseases. Similar to clinical practice, some works have demonstrated the benefits of multimodal fusion for automatic segmentation and classification using deep learning-based methods. However, current segmentation methods are limited to fusion of modalities with the same dimensionality (e.g., 3D+3D, 2D+2D), which is not always possible, and the fusion strategies implemented by classification methods are incompatible with localization tasks. In this work, we propose a novel deep learning-based framework for the fusion of multimodal data with heterogeneous dimensionality (e.g., 3D+2D) that is compatible with localization tasks. The proposed framework extracts the features of the different modalities and projects them into the common feature subspace. The projected features are then fused and further processed to obtain the final prediction. The framework was validated on the following tasks: segmentation of geographic atrophy (GA), a late-stage manifestation of age-related macular degeneration, and segmentation of retinal blood vessels (RBV) in multimodal retinal imaging. Our results show that the proposed method outperforms the state-of-the-art monomodal methods on GA and RBV segmentation by up to 3.10% and 4.64% Dice, respectively.

Exceeding the limit for microscopic image translation with a deep learning-based unified framework.

Deep learning algorithms have been widely used in microscopic image translation. The corresponding data-driven models can be trained by supervised or unsupervised learning depending on the availability of paired data. However, general cases are where the data are only roughly paired such that supervised learning could be invalid due to data unalignment, and unsupervised learning would be less ideal as the roughly paired information is not utilized. In this work, we propose a unified framework (U-Frame) that unifies supervised and unsupervised learning by introducing a tolerance size that can be adjusted automatically according to the degree of data misalignment. Together with the implementation of a global sampling rule, we demonstrate that U-Frame consistently outperforms both supervised and unsupervised learning in all levels of data misalignments (even for perfectly aligned image pairs) in a myriad of image translation applications, including pseudo-optical sectioning, virtual histological staining (with clinical evaluations for cancer diagnosis), improvement of signal-to-noise ratio or resolution, and prediction of fluorescent labels, potentially serving as new standard for image translation.

Differential Brain Activation for Four Emotions in VR-2D and VR-3D Modes.

Similar to traditional imaging, virtual reality (VR) imagery encompasses nonstereoscopic (VR-2D) and stereoscopic (VR-3D) modes. Currently, Russell's emotional model has been extensively studied in traditional 2D and VR-3D modes, but there is limited comparative research between VR-2D and VR-3D modes. In this study, we investigate whether Russell's emotional model exhibits stronger brain activation states in VR-3D mode compared to VR-2D mode. By designing an experiment covering four emotional categories (high arousal-high pleasure (HAHV), high arousal-low pleasure (HALV), low arousal-low pleasure (LALV), and low arousal-high pleasure (LAHV)), EEG signals were collected from 30 healthy undergraduate and graduate students while watching videos in both VR modes. Initially, power spectral density (PSD) computations revealed distinct brain activation patterns in different emotional states across the two modes, with VR-3D videos inducing significantly higher brainwave energy, primarily in the frontal, temporal, and occipital regions. Subsequently, Differential entropy (DE) feature sets, selected via a dual ten-fold cross-validation Support Vector Machine (SVM) classifier, demonstrate satisfactory classification accuracy, particularly superior in the VR-3D mode. The paper subsequently presents a deep learning-based EEG emotion recognition framework, adeptly utilizing the frequency, spatial, and temporal information of EEG data to improve recognition accuracy. The contribution of each individual feature to the prediction probabilities is discussed through machine-learning interpretability based on Shapley values. The study reveals notable differences in brain activation states for identical emotions between the two modes, with VR-3D mode showing more pronounced activation.

Improving drug response prediction via integrating gene relationships with deep learning.

Predicting the drug response of cancer cell lines is crucial for advancing personalized cancer treatment, yet remains challenging due to tumor heterogeneity and individual diversity. In this study, we present a deep learning-based framework named Deep neural network Integrating Prior Knowledge (DIPK) (DIPK), which adopts self-supervised techniques to integrate multiple valuable information, including gene interaction relationships, gene expression profiles and molecular topologies, to enhance prediction accuracy and robustness. We demonstrated the superior performance of DIPK compared to existing methods on both known and novel cells and drugs, underscoring the importance of gene interaction relationships in drug response prediction. In addition, DIPK extends its applicability to single-cell RNA sequencing data, showcasing its capability for single-cell-level response prediction and cell identification. Further, we assess the applicability of DIPK on clinical data. DIPK accurately predicted a higher response to paclitaxel in the pathological complete response (pCR) group compared to the residual disease group, affirming the better response of the pCR group to the chemotherapy compound. We believe that the integration of DIPK into clinical decision-making processes has the potential to enhance individualized treatment strategies for cancer patients.