Natural Vision Research Articles

Detection of Lane and Speed Breaker Warning System for Autonomous Vehicles Using Machine Learning Algorithm

Abstract: Road safety holds immense significance in today's world of transportation, primarily focusing on making roads safer to minimize the chances of accidents and injuries. Road safety revolves around creating and maintaining the road infrastructure. Our System enables vehicles to identify and slow down when approaching road bumps or speed breakers, and potholes. It utilizes cameras and sensors to help drivers stay within their designated lanes keeps an eye on lane markings and provides alerts if the vehicle begins to steer out of its lane, promoting safer and more reliable driving experiences. In high-risk areas or near accidentprone intersections, lane detection technology can trigger speed reduction measures such as illuminated signs and flashing lights when vehicles are not within their lanes. The integration of Natural Language processing (NLP), Computer Vision, Machine learning, and Deep learning can create a comprehensive approach to reducing accidents and improving overall road safety.

ProteinMAE: masked autoencoder for protein surface self-supervised learning.

The biological functions of proteins are determined by the chemical and geometric properties of their surfaces. Recently, with the booming progress of deep learning, a series of learning-based surface descriptors have been proposed and achieved inspirational performance in many tasks such as protein design, protein-protein interaction prediction, etc. However, they are still limited by the problem of label scarcity, since the labels are typically obtained through wet experiments. Inspired by the great success of self-supervised learning in natural language processing and computer vision, we introduce ProteinMAE, a self-supervised framework specifically designed for protein surface representation to mitigate label scarcity. Specifically, we propose an efficient network and utilize a large number of accessible unlabeled protein data to pretrain it by self-supervised learning. Then we use the pretrained weights as initialization and fine-tune the network on downstream tasks. To demonstrate the effectiveness of our method, we conduct experiments on three different downstream tasks including binding site identification in protein surface, ligand-binding protein pocket classification, and protein-protein interaction prediction. The extensive experiments show that our method not only successfully improves the network's performance on all downstream tasks, but also achieves competitive performance with state-of-the-art methods. Moreover, our proposed network also exhibits significant advantages in terms of computational cost, which only requires less than a tenth of memory cost of previous methods. https://github.com/phdymz/ProteinMAE.

Comparison of visual performance with iTrace analyzer following femtosecond laser-assisted cataract surgery with bilateral implantation of two different trifocal intraocular lenses.

To compare the postoperative binocular visual performance with an iTrace analyzer following femtosecond laser-assisted cataract surgery (FLACS) combined with bilateral implantation of two different types of diffractive trifocal intraocular lenses (IOL). During this retrospective observational study, patients who received bilateral FLACS combined with implantation of two different types of diffractive trifocal IOLs were evaluated. According to the IOLs' different types and design, the patients were divided into AT LISA tri839MP group (tri839 group) and AcrySof PanOptix TFNT00 group (TFNT group). Study parameters included preoperative and postoperative uncorrected distance visual acuity (UDVA) at 5 m, uncorrected near visual acuity (UNVA) at 30 cm and 40 cm, uncorrected intermediate visual acuity (UIVA) at 60 cm and 80 cm, postoperative refractive status, objective visual qualities and total high order aberrations (HOAs) postoperatively. The postoperative complications were also recorded. Totally 56 eyes of 28 patients (tri839 group, n=26; TFNT group, n=30) were included. Preoperative baseline characteristics between groups were not statistically significantly different. UDVA was not significantly different between groups except for 1wk follow-up due to the postoperative corneal edema. TFNT group showed statistically significant better UNIA at 60 cm than tri839 group at the 1wk (0.05±0.19 vs 0.15±0.10 logMAR, P=0.013), 1mo (0.05±0.12 vs 0.15±0.09 logMAR, P=0.001) and 3mo (0.04±0.12 vs 0.15±0.11 logMAR, P=0.001) follow-up, while tri839 group showed statistically significant better UNIA at 80 cm than TFNT group at the 1d (0.14±0.15 vs 0.20±0.14 logMAR, P=0.041) and 1mo (0.09±0.07 vs 0.14±0.10 logMAR, P=0.042) follow-up. Postoperative refractive status showed stable at every visit. Modulated transfer function (MTF) values and strehl ratio (SR) values were improved and HOAs were lower significantly after surgery. FLACS with bilateral implantations of both tri839 and TFNT00 can achieve satisfactory natural whole-course vision, high postoperative refractive stability and good visual quality but without significantly difference. iTrace aberration instrument can accurately evaluate the visual quality under different status.

Relphormer: Relational Graph Transformer for Knowledge Graph Representations

Transformers have achieved remarkable performance in widespread fields, including natural language processing, computer vision and graph mining. However, vanilla Transformer architectures have not yielded promising improvements in the Knowledge Graph (KG) representations, where the translational distance paradigm dominates this area. Note that vanilla Transformer architectures struggle to capture the intrinsically heterogeneous structural and semantic information of knowledge graphs. To this end, we propose a new variant of Transformer for knowledge graph representations dubbed Relphormer. Specifically, we introduce Triple2Seq which can dynamically sample contextualized sub-graph sequences as the input to alleviate the heterogeneity issue. We propose a novel structure-enhanced self-attention mechanism to encode the relational information and keep the semantic information within entities and relations. Moreover, we utilize masked knowledge modeling for general knowledge graph representation learning, which can be applied to various KG-based tasks including knowledge graph completion, question answering, and recommendation. Experimental results on six datasets show that Relphormer can obtain better performance compared with baselines.22Code is available in https://github.com/zjunlp/Relphormer.

Artificial Intelligence Using Python: Technologies and Applications

This article explores the significant role of Python in the development and application of artificial intelligence (AI). It covers fundamental AI concepts such as neural networks, machine learning, and deep learning, as well as key Python libraries used in this field. A brief overview of the history of AI is provided, starting from its early concepts and progressing to modern achievements. Python is regarded as the primary tool in AI development, and popular libraries and frameworks like TensorFlow, PyTorch, and scikit-learn are introduced. The article includes specific code examples and demonstrations of operations with Python libraries for machine and deep learning. It discusses AI application areas such as natural language processing, computer vision, recommendation systems, and robotics, along with examples of successful projects in these domains.

Interpretable Geometry Problem Solving Using Improved RetinaNet and Graph Convolutional Network

This paper proposes an interpretable geometry solution based on the formal language set of text and diagram. Geometry problems are solved using machines; however, machines encounter challenges in natural language processing and computer vision. Significant progress has improved existing methods in the extraction of geometric formal languages. However, the neglect of the graph structure information in the formal language and the lack of further refinement of the extracted language set can lead to poor theorem prediction and poor accuracy in problem solving. In this paper, a formal language graph is constructed using the extracted formal language set and applied to theorem prediction using a graph convolutional network. To better extract the relationship set of diagram elements, an improved diagram parser is proposed. The test results indicate that the improved method has good results when solving interpretable geometry problems.

Accelerating Attention Mechanism on FPGAs based on Efficient Reconfigurable Systolic Array

Transformer model architectures have recently received great interest in natural language, machine translation, and computer vision, where attention mechanisms are their building blocks. However, the attention mechanism is expensive because of its intensive matrix computations and complicated data flow. The existing hardware architecture has some disadvantages for the computing structure of attention, such as inflexibility and low efficiency. Most of the existing papers accelerate attention by reducing the amount of computation through various pruning algorithms, which will affect the results to a certain extent with different sparsity. This paper proposes the hardware accelerator for the multi-head attention (MHA) on field-programmable gate arrays (FPGAs) with reconfigurable architecture, efficient systolic array, and hardware-friendly radix-2 softmax. We propose a novel method called Four inputs Processing Element (FPE) to double the computation rate of the data-aware systolic array (SA) and make it efficient and load balance. Especially, the computation framework is well designed to ensure the utilization of SA efficiently. Our design is evaluated on a Xilinx Alveo U250 card, and the proposed architecture achieves 51.3×, 17.3× improvement in latency, and 54.4×, 17.9× energy savings compared to CPU and GPU.

Social Innovation in the Algorithmic Mind: Navigating the Psychological Landscape of Artificial Intelligence Reimagined

Artificial intelligence often termed as (AI) has emerged as a Changing force, revolutionizing various aspects of human life within the rapidly advancing realm of technology. While AI showcases immense potential for progress and innovation, its profound psychological impact on individuals and society cannot be overlooked. The paper further investigates the psychological impacts of AI across diverse domains, including education, healthcare, and social interaction. Offering insights into promising benefits and formidable barriers, the primary focus remains on the intricate nature of AI. In the swiftly evolving technological landscape, AI stands as a driving force with significant applications in various sectors such as education, healthcare, and manufacturing. Aligned with the National Strategy on Artificial Intelligence (NSAI) of India, AI encompasses systems capable of reasoning, learning, and adapting to changing environments, enhancing performance over time. Key components of AI, such as Machine Learning, Natural Language Processing (NLP), Computer Vision, Robotics, and Expert Systems, play pivotal roles in shaping its capabilities. While AI has the potential to address global challenges like climate change and diseases, responsible and ethical use is imperative to prevent the emergence of new issues, including social inequality and job displacement. This research paper provides a comprehensive exploration of the multifaceted psychological implications of AI, delving into its far-reaching effects on human cognition, emotion, behavior, and overall well-being.

Ferdynand Ruszczyc: A Polish Painter at the Crossroads of Cultures

The oeuvre of beloved Polish painter Ferdynand Ruszczyc (1870–1936) reflected the patriotic Neo-Romantic landscape trend of the fin-de-siècle prevalent in Germany and the Nordic countries (Denmark, Finland, Norway, Sweden). It should be considered in the context of Nordic visual culture for two reasons: (1) until the affiliation of Central and Eastern European nations with the Soviet Union in the wake of World War Two, nations bordering the Baltic formed a single, fluid territory of cultural exchange, and (2) Ruszczyc’s oeuvre displays significant commonalities with dominant patriotic and Neo-Romantic trends of progressive artists around the Baltic Sea, where landscape became a vehicle for expressing dreams and emotions, as well as love of homeland. This article situates Ruszczyc’s national and artistic identity at the crossroads of cultures and artistic impulses, regional as well as international. Ruszczyc was born in Bohdanów near Vilnius (now Belarus) to a Polish father and a Danish mother. Like many Polish artists from the Russian partition, he was educated at the Imperial Academy of Fine Arts in St. Petersburg, where he studied with Ivan Shishkin (1832–1898) and Arkhip Kuindzhi (1878–1910). He also travelled to Sweden. Ruszczyc was influenced by the Russian art circle Mir Iskusstva (World of Art, est. 1898) and is often compared with Nordic (e.g., Akseli Gallen-Kallela; Finnish, 1865–1931) and German (e.g., Otto Modersohn; 1865–1943) artists. His visions of nature are sometimes raw monumental images of the northern landscape or fairy-tale fantasies containing symbolic allusiveness and a mythical, poetic element that evoke intimate memories of the land of his childhood. In his paintings, Ruszczyc presented the changeability of seasons, orchards, soil and streams, clouds formations, and tree trunks with palpable emotion. By exposing the material substance of nature, his paintings also reveal its mystical aspect, its ability to transform in accordance with the cyclical, cosmic rhythm of growth, maturation, death, and rebirth.

Saccade execution increases the preview effect with faces: An EEG and eye-tracking coregistration study.

Under naturalistic viewing conditions, humans conduct about three to four saccadic eye movements per second. These dynamics imply that in real life, humans rarely see something completely new; there is usually a preview of the upcoming foveal input from extrafoveal regions of the visual field. In line with results from the field of reading research, we have shown with EEG and eye-tracking coregistration that an extrafoveal preview also affects postsaccadic visual object processing and facilitates discrimination. Here, we ask whether this preview effect in the fixation-locked N170, and in manual responses to the postsaccadic target face (tilt discrimination), requires saccade execution. Participants performed a gaze-contingent experiment in which extrafoveal face images could change their orientation during a saccade directed to them. In a control block, participants maintained stable gaze throughout the experiment and the extrafoveal face reappeared foveally after a simulated saccade latency. Compared with this no-saccade condition, the neural and the behavioral preview effects were much larger in the saccade condition. We also found shorter first fixation durations after an invalid preview, which is in contrast to reading studies. We interpret the increased preview effect under saccade execution as the result of the additional sensorimotor processes that come with gaze behavior compared with visual perception under stable fixation. In addition, our findings call into question whether EEG studies with fixed gaze capture key properties and dynamics of active, natural vision.

Depression Detection System

The Depression Detection System is an innovative technology designed to assist in the early detection and monitoring of depression. This system utilizes advanced machine learning algorithms and data analysis techniques to analyze various indicators andpatterns associated with depression, such as speech, facial expressions, and behavior. By continuously monitoring these signals, the system can identify potential signs of depression and provide valuable insights to healthcare professionals for accuratediagnosis and timely intervention. Major depressive disorder (MDD) or depression is among the most prevalent psychiatric disorders, affecting more than 300 million people globally. Early detection is critical for rapid intervention, which can potentially reduce the escalation of the disorder. The system operates by collecting data from various sources, including audio recordings of conversations, video recordings of facial expressions, and user- generated content from social media platforms. Through natural language processing and computer vision techniques, it extracts relevant features and patterns that may indicate depressive symptoms, such as changes in speech patterns, facial expressions of sadness or despair, or negative sentiment expressed in written content. Once the data is analyzed, the system generates comprehensive reports and visualizations, highlighting potential depressive indicators and their severity levels. These reports can assist healthcare providers in making informed decisions regarding further assessment and treatment planning. Additionally, the system can track the progression of depression over time, enabling clinicians to monitor the effectiveness of interventions and adjust treatment strategies accordingly. The Depression Detection System aims to complement traditional diagnostic methods by providing an objective and continuous assessment of depressive symptoms. By leveraging cutting-edge technology, this system has the potential to improve early detection rates, enhance patient outcomes, and facilitate timely interventions in the field of mental health. However, it is important to note that the system should always be used as a supportive tool and not as a replacement for professional clinical judgment and human interaction.

Integrity, Vulnerability, and Temporality

This paper asks how to account for vulnerable integrity in the temporal dynamism of human lives without relying on a subtractive vision of integral human nature, borrowing from presumed past or future rationality and maturity, or depending on an external attribution of dignity. Illustrating the challenges with vignettes from the author’s life, it argues inductively that human integrity includes morally inviolable vulnerability to others with whom we are in interdependent relationship and without whom we cannot develop or maintain our selves. Others reside at the core of our integrity, for better and for worse, and we reside at theirs. Augustine’s accounts of memory, time, and the narrative self; Whiteheadian process thought’s understanding of continuity through change; and feminist theories of narrative all provide theological and philosophical justifications for this vision of integrity. John Wall’s and Johan Brännmark’s non-foundational approaches to integrity and human rights lead us to the same conclusion without entailing theological anthropological claims, ensuring its relevance in a pluralist culture.

Securing AI‐based healthcare systems using blockchain technology: A state‐of‐the‐art systematic literature review and future research directions

AbstractHealthcare institutions are progressively integrating artificial intelligence (AI) into their operations. The extraordinary potential of AI is restricted by insufficient medical data for AI model training and adversarial attacks wherein attackers perturb the dataset by adding some noise to it, which leads to the malfunctioning of the AI models, and a lack of trust caused by the opaque operational approach it employs. This Systematic Literature Review (SLR) is a state‐of‐the‐art survey of the research on blockchain technology for securing AI‐integrated healthcare applications. The most relevant articles from the Scopus and Web of Science (WoS) databases were identified using the PRISMA model. Most of the existing literature is about protecting the healthcare data used by AI‐based healthcare systems using blockchain technology, but the modality of data (text, images, audio, and sound) was not specifically mentioned. Information on protecting the training phase and model deployment for AI‐based healthcare systems considering the variations in feature extraction based on the modality of data was also not clearly specified. Hence, the three subfields of AI, namely, natural language processing (NLP), computer vision, and acoustic AI are further studied to identify security loopholes in its implementation pipeline. The three phases, namely the dataset, the training phase, and the trained models need to be protected from adversaries to avoid malfunctioning of the deployed AI models. The nature of the data processed by NLP, computer vision, and acoustic AI, underlying deep neural network (DNN) architectures, the complexity of attacks, and the perceivability of attacks by humans are analyzed to identify the need for security. A blockchain solution for AI‐based healthcare systems is synthesized based on the findings that have demonstrated the distinctive technological features of blockchains. It offers a solution for the privacy and security issues encountered by NLP, computer vision, and acoustic AI to boost the widespread adoption of AI applications in healthcare.

Unleashing the Power of Deep Attention Networks: A Comprehensive Approach for Enhanced Artificial Intelligence

Deep learning has revolutionized the field of artificial intelligence by achieving state-of-the-art performance on a variety of complex tasks. Attention mechanisms have emerged as a powerful tool to enhance the capabilities of deep neural networks by enabling them to selectively focus on relevant information. In this article, we propose a novel artificial intelligence algorithm called Deep Attention Networks (DANs), which associate multiple attention mechanisms to improve performance on interesting tasks. We evaluate DANs on benchmark datasets in natural language processing, computer vision, and speech recognition and demonstrate superior results compared to existing state-of-the-art approaches. Our approach opens up new possibilities for advancing the field of artificial intelligence and holds promise for various real-world applications. Overall, our results demonstrate the effectiveness and potential of DANs for various AI applications, and highlight the power of combining deep neural networks with attention mechanisms.

Reflections on AI in Humanities

As artificial intelligence (AI) continues advancing rapidly, there is growing potential for its application in the humanities to uncover new insights and perspectives from the historical archives. However, it is also important to consider how AI tools themselves may unintentionally perpetuate existing biases if not developed conscientiously. This critical reflection reflects on the opportunities and challenges of utilising AI to amplify marginalised voices that have been traditionally excluded or underrepresented in mainstream historical narratives, with a focus on women. Through natural language processing and computer vision techniques, AI shows promise in automating the analysis of large volumes of text, image, and multimedia sources to bring to the surface female narratives previously overlooked due to limitations of manual research methods. However, issues such as training data bias, problematic stereotypes learned from legacy sources, and a lack of diversity among AI researchers threaten to replicate the very inequities they are seeking to overcome if not addressed proactively. Collaborative frameworks and design principles centred on representation, accountability and community oversight are needed. By critically examining its social responsibilities and impacts, this reflection argues that AI possesses great potential in the service of feminist and intersectional scholarship when guided appropriately. It calls for continued multidisciplinary dialogue to help ensure technologies amplify marginalised voices rather than risk their further marginalisation.

Coloring and fusing architectural sketches by combining a Y‐shaped generative adversarial network and a denoising diffusion implicit model

AbstractSketch‐based architectural design tasks require considerable time and effort. However, developing an artificial intelligence agent to color and fuse architectural sketches is challenging due to the imaginative nature of the task and technical limitations. Here, we first introduce Y‐shaped generative adversarial networks (GANs), which can color sketches on the basis of temporarily specified reference images by embedding an attention mechanism model that can match spatial and color information and by maintaining a type of triangular equilibrium state. Then, we develop practicable methods for coloring and fusing the given sketches simultaneously using a combination of our Y‐shaped GANs and the denoising diffusion implicit model, yielding results with high quality and veracity. A novel interpolation process is developed as a necessary trick for realizing the fusion function. The testing results indicate that this method works better than the existing methods for the tasks of coloring and fusing architectural sketches.

Deep neural network techniques for monaural speech enhancement and separation: state of the art analysis

AbstractDeep neural networks (DNN) techniques have become pervasive in domains such as natural language processing and computer vision. They have achieved great success in tasks such as machine translation and image generation. Due to their success, these data driven techniques have been applied in audio domain. More specifically, DNN models have been applied in speech enhancement and separation to perform speech denoising, dereverberation, speaker extraction and speaker separation. In this paper, we review the current DNN techniques being employed to achieve speech enhancement and separation. The review looks at the whole pipeline of speech enhancement and separation techniques from feature extraction, how DNN-based tools models both global and local features of speech, model training (supervised and unsupervised) to how they address label ambiguity problem. The review also covers the use of domain adaptation techniques and pre-trained models to boost speech enhancement process. By this, we hope to provide an all inclusive reference of all the state of art DNN based techniques being applied in the domain of speech separation and enhancement. We further discuss future research directions. This survey can be used by both academic researchers and industry practitioners working in speech separation and enhancement domain.

Investigations on artificially extending the spectral range of natural vision.

Organic semiconductors are being explored as retinal prosthetics with the prime attributes of bio-compatibility and conformability for seamless integration with the retina. These polymer-based artificial photoreceptor films are self-powered with light-induced signal strength sufficient to elicit neuronal firing events. The molecular aspect of these semiconductors provides wide spectral tunability. Here, we present results from a bulk heterostructure semiconductor blend with a wide spectral response range. This combination elicits clear spiking activity from a developing blind-chick embryonic retina in the subretinal configuration in response to white light. The response is largely triggered by the blue-green spectral regime rather than the red-NIR regime for the present polymer semiconductor layer attributes.

Research on transformer and attention in applied algorithms

The transformer is an encoder-decoder-based structure and model for deep learning that completely utilizes the self-attention mechanism. It has gained remarkable success in natural language processing and computer vision and is becoming the predominant research direction. This study first analyzes the transformer and attention mechanism, summarizes their advantages, and explores how they help the recommendation algorithm dynamically focus on specific parts of the input that are helpful to perform the current recommendation task. After analyzing the framework of the attention mechanism network and its weight computation for data received. To further enhance the practicality of objects in natural situations and the precision of object recognition, a transformer detection approach based on deformable convolution is presented. And analyzed how the transformer works in the generative pre-trained transformer. These algorithms illustrate the efficacy and robustness of the transformer, indicating that the transformer that incorporates the attention mechanism may satisfy the requirements of the majority of deep learning tasks. However, the unpredictability of demands, the exponential growth of information, and other issues will continue to make it challenging to deal with global interaction mechanisms and a unified framework for multimodal data.

Neural scaling of deep chemical models

Massive scale, in terms of both data availability and computation, enables important breakthroughs in key application areas of deep learning such as natural language processing and computer vision. There is emerging evidence that scale may be a key ingredient in scientific deep learning, but the importance of physical priors in scientific domains makes the strategies and benefits of scaling uncertain. Here we investigate neural-scaling behaviour in large chemical models by varying model and dataset sizes over many orders of magnitude, studying models with over one billion parameters, pre-trained on datasets of up to ten million datapoints. We consider large language models for generative chemistry and graph neural networks for machine-learned interatomic potentials. We investigate the interplay between physical priors and scale and discover empirical neural-scaling relations for language models in chemistry with a scaling exponent of 0.17 for the largest dataset size considered, and a scaling exponent of 0.26 for equivariant graph neural network interatomic potentials.