Goto

Collaborating Authors

Results


Color Channel Perturbation Attacks for Fooling Convolutional Neural Networks and A Defense Against Such Attacks

arXiv.org Artificial Intelligence

The Convolutional Neural Networks (CNNs) have emerged as a very powerful data dependent hierarchical feature extraction method. It is widely used in several computer vision problems. The CNNs learn the important visual features from training samples automatically. It is observed that the network overfits the training samples very easily. Several regularization methods have been proposed to avoid the overfitting. In spite of this, the network is sensitive to the color distribution within the images which is ignored by the existing approaches. In this paper, we discover the color robustness problem of CNN by proposing a Color Channel Perturbation (CCP) attack to fool the CNNs. In CCP attack new images are generated with new channels created by combining the original channels with the stochastic weights. Experiments were carried out over widely used CIFAR10, Caltech256 and TinyImageNet datasets in the image classification framework. The VGG, ResNet and DenseNet models are used to test the impact of the proposed attack. It is observed that the performance of the CNNs degrades drastically under the proposed CCP attack. Result show the effect of the proposed simple CCP attack over the robustness of the CNN trained model. The results are also compared with existing CNN fooling approaches to evaluate the accuracy drop. We also propose a primary defense mechanism to this problem by augmenting the training dataset with the proposed CCP attack. The state-of-the-art performance using the proposed solution in terms of the CNN robustness under CCP attack is observed in the experiments. The code is made publicly available at \url{https://github.com/jayendrakantipudi/Color-Channel-Perturbation-Attack}.


Over a Decade of Social Opinion Mining

arXiv.org Artificial Intelligence

Social media popularity and importance is on the increase, due to people using it for various types of social interaction across multiple channels. This social interaction by online users includes submission of feedback, opinions and recommendations about various individuals, entities, topics, and events. This systematic review focuses on the evolving research area of Social Opinion Mining, tasked with the identification of multiple opinion dimensions, such as subjectivity, sentiment polarity, emotion, affect, sarcasm and irony, from user-generated content represented across multiple social media platforms and in various media formats, like text, image, video and audio. Therefore, through Social Opinion Mining, natural language can be understood in terms of the different opinion dimensions, as expressed by humans. This contributes towards the evolution of Artificial Intelligence, which in turn helps the advancement of several real-world use cases, such as customer service and decision making. A thorough systematic review was carried out on Social Opinion Mining research which totals 485 studies and spans a period of twelve years between 2007 and 2018. The in-depth analysis focuses on the social media platforms, techniques, social datasets, language, modality, tools and technologies, natural language processing tasks and other aspects derived from the published studies. Such multi-source information fusion plays a fundamental role in mining of people's social opinions from social media platforms. These can be utilised in many application areas, ranging from marketing, advertising and sales for product/service management, and in multiple domains and industries, such as politics, technology, finance, healthcare, sports and government. Future research directions are presented, whereas further research and development has the potential of leaving a wider academic and societal impact.


AI, ML, 5G, IoT will be most important tech in 2021: Study

#artificialintelligence

Bengaluru, Nov 23: Artificial intelligence (AI), Machine learning, 5G and Internet of Things (IoT) would be the most important technologies in 2021, according to a new study by the Institute of Electrical and Electronics Engineers (IEEE). The technical professional organisation on Monday released the results of a survey of Chief Information Officers (CIO) and Chief Technology Officers (CTO) in the US, the UK, China, India and Brazil. The survey was on the most important technologies for 2021, the impact of the COVID-19 pandemic on the speed of their technology adoption and the industries expected to be most impacted by technology. On which would be the most important technologies, nearly one-third of the total respondents (32 per cent) said AI and ML followed by 5G (20 per cent) and IoT (14 per cent), according to an IEEE statement. Manufacturing (19 per cent), healthcare (18 per cent), financial services (15 per cent) and education (13 per cent) are the industries that most believe would be impacted by technology, according to the CIOs and CTOs surveyed.


How Can India Trump China In Higher Education Reforms For AI

#artificialintelligence

India has many milestones to achieve if it is to catch up with China, concluded a study published by the Observer Research Foundation, that offered a comparative analysis between the two in terms of higher education reforms for the development of talent in artificial intelligence (AI) and its research. The study, which compared various parameters including AI development plans and strategies of the two countries, their automation readiness index, talent retention, and research output, found China and India to have a very evident difference in their approach to bring reforms in higher education for AI. "Countries' strength in AI will potentially impact their position in the power structure in international politics in the long term," said Dr Romi Jain, author of the study and a Postdoctoral Research Fellow at the University of British Columbia. "From the geopolitical perspective, acquisition of robust AI capabilities could even shape the contours and outcomes of Sino-Indian conflict, competition and rivalry." As both countries introduce AI degree programmes in higher education, the results are too premature to be judged.


News from a postpandemic world

Science

We asked young scientists to imagine this scenario: You are a science writer in the year 2040 working on a news story that answers this question: What do you hope or fear will be the long-term effects of the coronavirus disease 2019 (COVID-19) pandemic? A selection of their responses, arranged as a newpaper, is below. Follow NextGen Voices on Twitter with hashtag #NextGenSci. Read previous NextGen Voices survey results at . —Jennifer Sills Today, scientists confirm that 1000 previously endangered species have been removed from the Vulnerable list. Biodiversity renewal has been under way since the COVID-19 pandemic 20 years ago led many governments to reevaluate their priorities. Hunting practices and bushmeat consumption were constrained to limit the transmission of new pathogens through human contact with the meat and biofluids of wild animals. Deforestation was restricted worldwide when it became clear that land-use modifications and climate change were important drivers of vector-borne diseases. COVID-19 claimed many lives, but the political and environmental changes the pandemic inspired have likely saved many more by protecting the world's biodiversity. Joel Henrique Ellwanger Department of Genetics, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, 91501-970, Brazil. Email: joel.ellwanger{at}gmail.com Science and technology research budgets, now classified as an arm of the national defense force, could rival traditional military spending in a few years' time. This newfound prioritization of science was shaped by the COVID-19 pandemic, which made clear that the previous conception of military force is impractical when the enemy is invisible and formidable. The unprecedented redirection of financial resources to scientific communities to help find a cure and vaccines, along with the increased demand for scientific experts, expanded technological frontiers and gave science a well-deserved space in governance. Mpho Diphago Stanley Lekgoathi The South African Nuclear Energy Corporation, Pretoria, Gauteng, South Africa. Email: mpho.lekgoathi{at}necsa.co.za In response to the 50th wave of COVID-19, which hit New York City last month, the U.S. government has announced that the first spaceship designated for in-orbit medical treatment of COVID-19 patients will soon transport 10,000 residents from high-risk zones to Space. Scientists say that prolonged stay in Space colonies with exposure to controlled gamma radiation from cosmic dust may help weaken the virus's strong affinity to lung tissue. “We will do all we can to protect our residents on Earth. Unlike 2019, we are prepared for this challenge,” said the President in a Capitol Hill address. The Senate has voted to fund the treatment expenses for everyone on the flight. Kartik Nemani Layered Materials and Structures Lab, Department of Mechanical and Energy Engineering, Purdue School of Engineering and Technology, Indiana University–Purdue University Indianapolis, Indianapolis, IN 46202, USA. Email: snemani{at}purdue.edu Workers at major corporations staged a walk-out today, the 20th anniversary of the COVID-19 pandemic, to protest what some have deemed invasive monitoring. Many fears subsided when the first SARS-CoV-2 vaccine was broadly distributed in 2023, but subsequent zoonotic viruses emerged faster than society could prepare for them. With the world economy precariously weak, a cautious arrangement was reached: Workers could return to their jobs if they submitted to routine infection checks. At first, these were relatively innocuous temperature probes and cough tracking. However, with the 2029 advent of low-cost RNA wastewater screening by smart toilets and ubiquitous wall-mounted infrared heat sensors, infected employees could be pinpointed before displaying acute symptoms. Later, an eCommerce/fitness-tracking consortium released artificial intelligence algorithms that combined smartwatch health metrics and recent online search history. Corporate Wellness Boards used the results to justify mandatory quarantines. Employees cried foul. The debate rages on in our courts and on the Giganet about whether the public good is served by exposing the “viral status” of the few. Michael A. Tarselli Society for Laboratory Automation and Screening, Oak Brook, IL 60523, USA. Email: mtarselli{at}slas.org Earlier this month, 21 individuals were quarantined in Kampala, Uganda, after a man was diagnosed with Marburg hemorrhagic fever by the local laboratory of the International Center for Disease Prevention (ICDP). The patient, who has now fully recovered, may have been infected at the veterinary clinic where he worked in close contact with possible animal carriers. “This is a virus that spreads easily through bodily fluids and historically has been transmitted to caregivers,” said Dr. Icuaf, director of the ICDP. Once again, the localized presence of centers with efficient testing capabilities made it possible to identify patient zero and contain the outbreak at its inception. As a result, “no deaths occurred, and everyone who might have been exposed has been quarantined while we monitor their health,” added Dr. Icuaf. The ICDP was instituted in 2021 as a global response to the COVID-19 pandemic, which marked a revolution in public awareness of science-based policy. The cost of crisis prevention is now routinely compared with the predicted price of managing such a crisis after it has occurred. Ahmed Al Harraq Cain Department of Chemical Engineering, Louisiana State University, Baton Rouge, LA 70803, USA. Email: aahme22{at}lsu.edu One of the world's leading universities is launching a large-scale screen of potential antiviral and antibacterial drugs on human volunteers. The substances show promising results in vitro but have not been tested on animals. To compensate for the risk of side effects, all volunteers will receive generous payment. “Drugs showing promising effects on mice could be ineffective on humans, making drug development expensive and slow,” explained the leading scientist of the drug screen. Human rights experts warned against granting permission to conduct the study. “Offering payment for causing physical harm targets the economically vulnerable and violates basic human rights,” they argued. However, doctors and politicians praise the idea, referring to the COVID-19 epidemic. “Developing a new drug through the traditional process can take years. Testing multiple potential candidates on coronavirus-infected people saved thousands of lives before basic research had a chance to catch up. Next time, we want to be prepared,” explained the health minister. Anna Uzonyi Department of Molecular Genetics, Weizmann Institute of Science, Rehovot, 7610001, Israel. Email: anna.uzonyi{at}weizmann.ac.il Results published today from a 20-year experiment show that a “lottery” grant funding scheme is superior to traditional peer-review assessment panels. For decades, researchers have debated the effectiveness and cost-efficiency of selecting grant recipients through a peer-review process, given the documented biases that hinder diversity and equitable decision-making. “It was a controversial move at the time, but the results are clear,” said the lead author of the study. The funding experiment, which began in 2020 in response to the COVID-19 pandemic, was introduced to preserve the workforce employed on short-term contracts. During that year, pandemic-related budget cuts and social restrictions impeded the traditional peer-review process. “The lottery not only reduced peer-review bias but also added millions of dollars per year to the sector in hours saved by academics no longer devoting time to peer review,” said the lead author. “That time was spent on doing more experiments, mentoring colleagues, or achieving a healthier work-life balance.” Ken Dutton-Regester Department of Genetics and Computational Biology, QIMR Berghofer Medical Research Institute, Brisbane, QLD 4006, Australia. Twitter: @stemventurist As the debate continues on the efficacy of educational methods, most universities now use a combination of in-person, remote, and technology-enhanced classrooms. The rapid expansion of evidence-based strategies such as machine learning and artificial intelligence, audio and video tools, three-dimensional environments, and simulations across disciplines began during the COVID-19 pandemic. The decision to move education to a computer-based environment to protect the health and safety of students and staff transformed the educational conversation. In the increasingly technology-enhanced world, discussions about how to teach a science class online, how to facilitate lab experiences, and how to conduct experiments with new constraints swept the research community. A nuanced understanding emerged about true online pedagogy versus synchronous, remote meetings. Two decades later, we see the results of this transformation. Rachel Yoho Department of Environmental and Global Health, University of Florida, Gainesville, FL 32603, USA. Twitter: @rachel_yoho A stunning 200,000 people attended the grand opening ceremony of the 2040 Olympics yesterday in New Delhi, India. It has been 20 years since such a public event could take place safely. Only with the recent release of clothing and shoes made of technologically advanced materials that instantly kill viruses could the social distancing that began with the COVID-19 pandemic be relaxed. For added peace of mind, all attendees at the ceremony consented to the skin implantation of Viroclean, a new chip-based device that sounds an alarm when it detects viruses in the air. Sudhakar Srivastava Institute of Environment and Sustainable Development, Banaras Hindu University, Varanasi, Uttar Pradesh, 221005, India. Email: sudhakar.srivastava{at}gmail.com This weekend, at the Coachella 2040 music festival, three aerosol biosurveillance sensors detected a SARS-like virus in the air. Smartphone tracing, using the opt-in U.S. Centers for Disease Control and Prevention (CDC) geospatial health app developed in the wake of the COVID-19 pandemic, identified two potential index cases. The CDC outbreak prevention team mobilized regional contact tracers to intercept and test both individuals within an hour of first detection. One individual tested positive for a variant of the 2019 SARS-CoV-2 strain, previously thought to be eradicated, and is undergoing treatment in quarantine. Michael Strong Center for Genes, Environment, and Health, National Jewish Health and University of Colorado, Anschutz Medical Campus, Denver, CO 80206, USA. Email: strongm{at}njhealth.org Last week's 15th annual Pan-global Remote Integrated Sciences Meeting (PRISM) attracted more than 100,000 attendees from more than 160 countries. Scientists, educators, students, entrepreneurs, policymakers, and industry experts from fields spanning the physical, biological, and social sciences logged on to the online venue, enabled by virtual reality. Advanced machine learning algorithms provided recommendations for presentations relevant to each participant based on both their expertise and potential for interdisciplinary collaboration. As usual, the highlight of the meeting was the virtual poster sessions, driven by interactivity and streamlined to optimize small-group scientific conversation across fields. Many junior scientist attendees were surprised to learn that such events were nearly unheard of before PRISM grew from the increasing move toward virtual conferences during the coronavirus pandemic over 20 years ago. “My adviser told me that when she was a grad student, big conferences were all held in person,” writes one anonymous Ph.D. student. “Can you imagine having a giant conference like this in some random convention center, with tens of thousands of scientists spending hundreds of dollars on fuel-inefficient flights and hotel booking, lugging around printed posters and just milling around for a week trying to find the optimal talks to attend? Insane.” Yifan Li Department of Chemistry, University of California, Berkeley, Berkeley, CA 94720, USA. Twitter: @iWonderWhyly Today, cell-based meat consumption has surpassed farm-produced meat for the first time. The transition began with the meat shortages and near collapse of the meat supply chain during the COVID-19 outbreak. With thousands of workers packed into poorly ventilated and unhygienic facilities, meat processing plants were hotspots for the SARS-CoV-2 virus. A global meat shortage emerged as production rates were slashed. Most people turned to the plant-based meat alternatives available at the time. The meat industry's demise was sealed when cell-based meat entered the mainstream market the following year. Clean meat eliminated the negative effects of the meat industry, from pollution caused by runoff and antibiotics, to worker and animal cruelty, to the carbon footprint of livestock, which contributed 18% of greenhouse gas emissions at the time. Cell-based meat has been growing in popularity ever since, as traditional meat became ethically and environmentally unpalatable. JiaJia Fu Whittle School and Studios, Washington, DC 20008, USA. Email: jjnaturalist{at}gmail.com Global seafood supply now relies entirely on aquaculture. The turning point came when researchers optimized the breeding techniques for edible crabs, enabling high-valued crab species such as mud crabs and blue crabs to be mass-produced in full aquaculture settings. The prioritization of aquaculture was made possible by the COVID-19 pandemic in 2020. A 12-month closure of fisheries during the wave of global stay-at-home orders led to the rejuvenation of overexploited species such as sardines and mackerels, which had been on the verge of extinction, and made people recognize the fragility of the supply chain. Full investment in aquaculture research began the following year. Khor Waiho Institute of Tropical Aquaculture and Fisheries, Universiti Malaysia Terengganu, Kuala Nerus, Terengganu, 21030, Malaysia. Email: waiho{at}umt.edu.my Next week, the United Nations will meet to assess whether the goals of the 2040 Agenda for Sustainable Development have been achieved. Unfortunately, reasons for optimism are scarce. Overexploitation of natural resources, CO2 emissions, and plastic waste continue to soar. The wealthiest sector of the population consumes 80% of the resources, and the poorest people increasingly suffer from extreme weather events, famines, and freshwater scarcity. We were already heading in this direction early in the century, when the short-term vision of corporations and policy-makers prioritized economic benefits over human and environmental health. The COVID-19 pandemic exacerbated the negative trends. Since 2020, an array of wasteful practices increased, including the proliferation of single-use products and travel in private vehicles to avoid physical contact. After reviewing the past decade, the UN countries will discuss commitments to decrease inequality and pollution by 2050. Isabel Marín Beltrán Laboratory of Environmental Technologies, Centro de Ciências do Mar do Algarve, Universidade do Algarve, Campus de Gambelas, Faro, 8005-139, Portugal. Email: imbeltran{at}ualg.pt For the first time, global average air temperature is more than 2°C higher than the 20th-century global average. Scientists suggest that decisions made in response to the COVID-19 pandemic led to today's disastrous climate consequences. After the Paris Climate Agreement in 2016, scientists were hopeful. National governments were implementing increasingly ambitious measures to meet their commitments. But the economic fallout of the pandemic led growing economies such as India to relax environmental clearance guidelines for industries and infrastructure projects and cut funding allocated to environmental reforms. First-world countries such as the United States and China, instead of shifting toward renewable energy, boosted investment in fossil fuels, which in turn increased greenhouse gas emissions. Even after multiple warnings from the Intergovernmental Panel on Climate Change, G20 nations neglected to follow the advice of scientists. Akash Mukherjee Department of Physics, Indian Institute of Science Education and Research, Pune, Pune, Maharashtra, 411008, India. Twitter: @aghori_AM A government report released yesterday warns of a potential spike in counterfeit immunity passports entering the market this coronavirus season. According to Jane London, the U.K. health minister, “There is a substantial increase in the number of illegal immunigrants crossing provincial and municipal borders. The public should be aware that just scanning someone's immunity passport is not enough anymore.” This report comes just 6 months after the U.S. Centers for Disease Control and Prevention first released notice that the “NextGen Immunity Passport” brand had been hacked, allowing scammers and tech-savvy citizens to falsify the immunity data they carry with them by law. Asked how businesses and town-guards were detecting falsified immunity passports at checkpoints, minister of national movement John Petersfield told journalists, “This is a police matter. Any further information about detection at this time will only help counterfeiters.” Widespread counterfeiting, as well as last year's false-negative scandal, has generated substantial public distrust in the use of the immunity passport system in movement legislation, now 19 years old. “We learned our lesson about free movement back in 2020,” said one government official who wished to remain anonymous, “but the immunity passport system is cracking, and we don't see a fix yet.” Tyler D. P. Brunet Department of History and Philosophy of Science, University of Cambridge, Cambridge, Cambridgeshire, CB2 3RH, UK. Email: tdpb2{at}cam.ac.uk


AI helps spot early signs of glaucoma progression to blindness - CRN - India

#artificialintelligence

Using Artificial Intelligence (AI), researchers have developed a quick test to identify which people with glaucoma are at risk of rapid progression to blindness. A new test can detect glaucoma progression 18 months earlier than the current gold standard method, said the study published in the journal Expert Review of Molecular Diagnostics. Glaucoma, the leading global cause of irreversible blindness, affects over 60 million people, which is predicted to double by 2040 as the global population ages. Loss of sight in glaucoma is caused by the death of cells in the retina, at the back of the eye. "Being able to diagnose glaucoma at an earlier stage, and predict its course of progression, could help people to maintain their sight, as treatment is most successful if provided at an early stage of the disease," said study first author Eduardo Normando from Imperial College London.


Impact of different belief facets on agents' decision -- a refined cognitive architecture

arXiv.org Artificial Intelligence

This paper presents a conceptual refinement of agent cognitive architecture inspired from the beliefs-desires-intentions (BDI) and the theory of planned behaviour (TPB) models, with an emphasis on different belief facets. This enables us to investigate the impact of personality and the way that an agent weights its internal beliefs and social sanctions on an agent's actions. The study also uses the concept of cognitive dissonance associated with the fairness of institutions to investigate the agents' behaviour. To showcase our model, we simulate two historical long-distance trading societies, namely Armenian merchants of New-Julfa and the English East India Company. The results demonstrate the importance of internal beliefs of agents as a pivotal aspect for following institutional rules.


A Time Series Analysis-Based Stock Price Prediction Using Machine Learning and Deep Learning Models

arXiv.org Machine Learning

Prediction of future movement of stock prices has always been a challenging task for the researchers. While the advocates of the efficient market hypothesis (EMH) believe that it is impossible to design any predictive framework that can accurately predict the movement of stock prices, there are seminal work in the literature that have clearly demonstrated that the seemingly random movement patterns in the time series of a stock price can be predicted with a high level of accuracy. Design of such predictive models requires choice of appropriate variables, right transformation methods of the variables, and tuning of the parameters of the models. In this work, we present a very robust and accurate framework of stock price prediction that consists of an agglomeration of statistical, machine learning and deep learning models. We use the daily stock price data, collected at five minutes interval of time, of a very well known company that is listed in the National Stock Exchange (NSE) of India. The granular data is aggregated into three slots in a day, and the aggregated data is used for building and training the forecasting models. We contend that the agglomerative approach of model building that uses a combination of statistical, machine learning, and deep learning approaches, can very effectively learn from the volatile and random movement patterns in a stock price data. We build eight classification and eight regression models based on statistical and machine learning approaches. In addition to these models, a deep learning regression model using a long-and-short-term memory (LSTM) network is also built. Extensive results have been presented on the performance of these models, and the results are critically analyzed.


AI by the numbers

#artificialintelligence

Artificial intelligence is having a significant impact on mainstream business and computing after years of being in the hype cycle. Companies such as Amazon and Netflix have saved billions of dollars a year, and AI is expected to boost the global economy by trillions of dollars over the next several years. Concerns about AI's use remain, however, including security risks and biases it could introduce into hiring and society as a whole as well as bad decisions it might make due to poor underlying data quality. Here's a snapshot of the present and future of AI, told in 11 statistics: That's a 14 percent increase, more than the current economic output of China and India combined, a PwC study projects. Some $6.6 trillion of the boost will come from increased productivity, while $9.1 trillion will arrive as a result of increased economic consumption.


Strategies to Tackle the Global Burden of Diabetic Retinopathy: From Epidemiology to Artificial Intelligence

#artificialintelligence

Diabetes is a global public health disease projected to affect 642 million adults by 2040, with about 75% residing in low- and middle-income countries. Diabetic retinopathy (DR) affects 1 in 3 people with diabetes and remains the leading cause of blindness in working-aged adults. There are 3 broad strategic imperatives to prevent blindness caused by DR. Primary prevention requires preventing or delaying the onset of DR in those with diabetes by systems-level lifestyle modifications such as increasing physical activity or dietary modifications, pharmacological interventions for glycaemic and blood pressure control, and systematic screening for the onset of DR. Secondary prevention requires preventing the progression of DR in patients with DR by continuing systemic risk factor control, regular screening to monitor for the progression of mild DR to vision-threatening stages, and the development and implementation of evidence-based guidelines for managing DR. In this aspect, telemedicine-based DR screening incorporating artificial intelligence technology has the potential to facilitate more widespread and cost-effective screening, particularly in low- and middle-income countries. Tertiary prevention of DR blindness has been the main focus of the clinical ophthalmology community, classically based on laser photocoagulation treatment and ocular surgery but with an increasing use of anti-vascular endothelial growth factor (anti-VEGF) for vision-threatening DR. Evidence from serial epidemiological studies shows blindness due to DR has declined in high-income countries (e.g., the USA and UK) due to coordinated public health education efforts, increased awareness, early detection by DR screening, sustained systemic risk factor control, and the availability of effective tertiary level treatment. However, the progress made in reducing DR blindness in high-income countries may be overwhelmed by the increasing numbers of patients with diabetes and DR in low- and middle-income countries (e.g., China, India, Indonesia, etc.).