Category Archives: Machine Learning
Beyond Algorithms: How AI is Learning Our Social Cues – DataDrivenInvestor
The journey of artificial intelligence (AI) has been nothing short of remarkable. From its inception in the mid-20th century, AI was envisioned as a means to mimic human intelligence. This vision was rooted in the belief that machines could be programmed to perform tasks that typically require human cognition. The early years of AI were characterized by optimism and a focus on creating systems that could solve logical problems and perform specific, rule-based tasks.
Initially, AIs triumphs were in areas that demanded computational prowess rather than emotional intelligence. For instance, the world witnessed AIs potential when IBMs Deep Blue defeated chess grandmaster Garry Kasparov in 1997. These early achievements, though impressive, were confined to the realms of mathematics and logic. They demonstrated AIs ability to process and execute complex algorithms but did not venture into the nuances of human emotions or social behaviours.
As technology progressed, so did the capabilities of AI. The focus shifted from performing rudimentary, rule-based tasks to tackling more complex activities. This transition was marked by the advent of machine learning a branch of AI that learns from and makes decisions based on data.
Enabling AI to interpret social cues is fraught with challenges. The world of human emotion and social interaction is rich, complex, and often subjective. Teaching a machine to navigate this world involves not just technological hurdles but also ethical and cultural considerations.
Machine learning, along with natural language processing (NLP) and computer vision, became instrumental in evolving AI from a tool of computational logic to one capable of understanding and interacting with the human world in a more nuanced way.
Today, AI stands on the brink of a new frontier: social intelligence. This emerging domain represents a significant leap from traditional AI capabilities. Social intelligence in AI refers to the ability of machines to understand and appropriately respond to human social cues such as facial expressions, tone of voice, body language, and contextual subtleties. This development is not just a technological achievement but a bridge towards more empathetic and effective human-machine interactions.
Data Acquisition
AIs journey in understanding human interaction begins with data acquisition. This involves collecting a vast array of social data, such as text (from social media, emails, chat conversations), speech (voice recordings, call center data), visual cues (videos, images capturing facial expressions and body language), and even physiological signals (like heart rate or skin conductance). The quality and diversity of this data are crucial for the accuracy and comprehensiveness of social cue interpretation.
Alongside NLP, developments in computer vision, particularly in facial recognition, opened new avenues for AI in social understanding. AI systems began to recognize and interpret human facial expressions, a fundamental aspect of non-verbal communication. Emotion analysis algorithms were developed, allowing AI to infer emotions based on facial cues, a step closer to mimicking human empathy and understanding.
In this evolving landscape, optimism abounds. As AI ethicist Kate Darling remarks,
AI can unlock new possibilities we cannot yet envision.
With responsible research, development, and collaboration across disciplines, AI systems can gain social nuance and adaptability. The promise of a future where AI understands and augments our social interactions is within reach.
Follow me on LinkedIn for updates on AI Trends
Go here to see the original:
Beyond Algorithms: How AI is Learning Our Social Cues - DataDrivenInvestor
NIST Identifies Types of Cyberattacks That Manipulate Behavior of AI Systems | NIST – NIST
An AI system can malfunction if an adversary finds a way to confuse its decision making. In this example, errant markings on the road mislead a driverless car, potentially making it veer into oncoming traffic. This evasion attack is one of numerous adversarial tactics described in a new NIST publication intended to help outline the types of attacks we might expect along with approaches to mitigate them.
Credit: N. Hanacek/NIST
Adversaries can deliberately confuse or even poison artificial intelligence (AI) systems to make them malfunction and theres no foolproof defense that their developers can employ. Computer scientists from the National Institute of Standards and Technology (NIST) and their collaborators identify these and other vulnerabilities of AI and machine learning (ML) in a new publication.
Their work, titled Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations (NIST.AI.100-2), is part of NISTs broader effort to support the development of trustworthy AI, and it can help put NISTs AI Risk Management Framework into practice. The publication, a collaboration among government, academia and industry, is intended to help AI developers and users get a handle on the types of attacks they might expect along with approaches to mitigate them with the understanding that there is no silver bullet.
We are providing an overview of attack techniques and methodologies that consider all types of AI systems, said NIST computer scientist Apostol Vassilev, one of the publications authors. We also describe current mitigation strategies reported in the literature, but these available defenses currently lack robust assurances that they fully mitigate the risks. We are encouraging the community to come up with better defenses.
AI systems have permeated modern society, working in capacities ranging from driving vehicles to helping doctors diagnose illnesses to interacting with customers as online chatbots. To learn to perform these tasks, they are trained on vast quantities of data: An autonomous vehicle might be shown images of highways and streets with road signs, for example, while a chatbot based on a large language model (LLM) might be exposed to records of online conversations. This data helps the AI predict how to respond in a given situation.
One major issue is that the data itself may not be trustworthy. Its sources may be websites and interactions with the public. There are many opportunities for bad actors to corrupt this data both during an AI systems training period and afterward, while the AI continues to refine its behaviors by interacting with the physical world. This can cause the AI to perform in an undesirable manner. Chatbots, for example, might learn to respond with abusive or racist language when their guardrails get circumvented by carefully crafted malicious prompts.
For the most part, software developers need more people to use their product so it can get better with exposure, Vassilev said. But there is no guarantee the exposure will be good. A chatbot can spew out bad or toxic information when prompted with carefully designed language.
In part because the datasets used to train an AI are far too large for people to successfully monitor and filter, there is no foolproof way as yet to protect AI from misdirection. To assist the developer community, the new report offers an overview of the sorts of attacks its AI products might suffer and corresponding approaches to reduce the damage.
The report considers the four major types of attacks: evasion, poisoning, privacy and abuse attacks. It also classifies them according to multiple criteria such as the attackers goals and objectives, capabilities, and knowledge.
Evasion attacks, which occur after an AI system is deployed, attempt to alter an input to change how the system responds to it. Examples would include adding markings to stop signs to make an autonomous vehicle misinterpret them as speed limit signs or creating confusing lane markings to make the vehicle veer off the road.
Poisoning attacks occur in the training phase by introducing corrupted data. An example would be slipping numerous instances of inappropriate language into conversation records, so that a chatbot interprets these instances as common enough parlance to use in its own customer interactions.
Privacy attacks, which occur during deployment, are attempts to learn sensitive information about the AI or the data it was trained on in order to misuse it. An adversary can ask a chatbot numerous legitimate questions, and then use the answers to reverse engineer the model so as to find its weak spots or guess at its sources. Adding undesired examples to those online sources could make the AI behave inappropriately, and making the AI unlearn those specific undesired examples after the fact can be difficult.
Abuse attacks involve the insertion of incorrect information into a source, such as a webpage or online document, that an AI then absorbs. Unlike the aforementioned poisoning attacks, abuse attacks attempt to give the AI incorrect pieces of information from a legitimate but compromised source to repurpose the AI systems intended use.
Most of these attacks are fairly easy to mount and require minimum knowledge of the AI system and limited adversarial capabilities, said co-author Alina Oprea, a professor at Northeastern University. Poisoning attacks, for example, can be mounted by controlling a few dozen training samples, which would be a very small percentage of the entire training set.
The authors who also included Robust Intelligence Inc. researchers Alie Fordyce and Hyrum Anderson break down each of these classes of attacks into subcategories and add approaches for mitigating them, though the publication acknowledges that the defenses AI experts have devised for adversarial attacks thus far are incomplete at best. Awareness of these limitations is important for developers and organizations looking to deploy and use AI technology, Vassilev said.
Despite the significant progress AI and machine learning have made, these technologies are vulnerable to attacks that can cause spectacular failures with dire consequences, he said. There are theoretical problems with securing AI algorithms that simply havent been solved yet. If anyone says differently, they are selling snake oil.
Follow this link:
NIST Identifies Types of Cyberattacks That Manipulate Behavior of AI Systems | NIST - NIST
How helpful was artificial intelligence to online retailers in 2023? – Digital Commerce 360
With more than a full year in the books for generative AI platforms like ChatGPT and DALL-E, retailers have had time to test the technology and see what it looks like when implemented into their businesses. But generative is just part of the equation. The larger, far more expansive field of artificial intelligence entered standard operating procedures in various forms for many online retailers in 2023.
Between artificial intelligence and machine learning, automation in retail is becoming increasingly common. Whereas artificial intelligence refers to technology that can mimic human intelligence, machine learning is different. Machine learning technology enables a program to perform specific tasks and provide accurate results by identifying patterns. And as online retailers continue to use both, the lines can sometimes get blurry, but the results are clear.
Salesforce said artificial intelligence accounted for $194 billion in online holiday sales, primarily through predictive recommendations. And thats just in November and December. The software provider said artificial intelligence influenced 17% of all online orders in the last two months of 2023.
Below, we recap some of Digital Commerce 360s most insightful coverage about artificial intelligence (including generative AI) and machine learning in online retail from the past year. These stories highlight meaningful AI/ML trends among online retailers in 2023. Most notably, they include use cases spanning from product design to chatbots and digital marketing, and much more.
Online pet retailer Finn invested in artificial intelligence to appeal to specific groups of customers quickly.
SodaStream invested in artificial intelligence to determine which ad campaigns would be most successful via email, SMS text and on social media.
Thousands of shoppers each month negotiate with Industry Wests artificial intelligence chatbot in hopes of reaching a deal for a product discount.
How SMBs are using AI
Small and medium-sized businesses like mens grooming retailer Huron are using AI to balance financials easily. The retailer is also balancing how it sells to customers shopping via Amazon versus its direct-to-consumer website. The brand is using plug-ins to upsell.
Googles AI dressing room technology could reduce ecommerce returns and give retailers data to better target consumers, experts say.
Tailored Brands Inc. invests in artificial intelligence to understand its retail and rental customers for digital and in-store shoppers.
Online music instrument and equipment retailer Sweetwater increases email open rates and online sales thanks to AI-generated email recommendations.
Machine learning software enables Mars Petcare to measure how appealing pet food images are to online consumers.
The My Skin Biome tool from Beekman 1802 and Perfect Corp. works directly from the website on a users mobile phone.
Generative AI has been a key discussion topic all year. Online retailers are already incorporating it into their design processes to come up with new products and variations of existing products.
Adding shoppable products on both English and Spanish blog posts which are AI-generated has helped the retailer more than triple its average order value.
Menswear retailer Otero attributes its low return rate to the accuracy of its online fitting tool using Perfitly.
Impressed with the sophistication of generative AI chatbots, ski and sporting goods brand Evo plans to launch a customer service chatbot in time for the holiday season.
The generative AI tool creates and publishes a summary of all of the reviews at the top of the customer reviews section on the product detail page.
Large tool manufacturer Stanley Black & Decker is looking for a generative AI tool to write product descriptions and speed up its product detail product optimization. But the technology is not there yet.
Generative AI systems like ChatGPT are the hottest thing in tech these days, and some retailers will be showing off the power of the technology during the upcoming holiday season.
Toothpaste CPG brand Colgate-Palmolive tests a generative AI chatbot to more efficiently gather analysis to create better-converting product detail pages.
Generative AI is a valuable tool for digital marketers looking to simplify tasks. The technology is a creative reservoir for good and bad and sometimes outright goofy ideas. But thats a good thing when trying to stand out from the competition, retailers say.
Major retailers and consumer brands including eBay, Colgate, Ghirardelli, Newegg and Stanley Black & Decker are using generative AI today to speed product detail page content creation or optimization. While some have AI-created content live today, others are still perfecting their tools before debuting them to the public.
The chocolate brands ecommerce content operations and development manager shares how the major chocolate brand is using AI to make decisions about product detail page images.
Queenly uses generative AI to populate product listings from a series of questions answered by online resellers.
Submit your data withthis quick surveyand well see where you fit in our next ranking update.
Stay on top of the latest developments in the ecommerce industry. Sign up for a complimentary subscription toDigital Commerce 360 Retail News.Follow us onLinkedIn,TwitterandFacebook. Be the first to know when Digital Commerce 360 publishes news content.
Read the original:
How helpful was artificial intelligence to online retailers in 2023? - Digital Commerce 360
Machine learning identifies promising antibacterial ruthenium-based drug candidates – Chemistry World
A machine learning model has been created that can identify ruthenium-based antibiotic drug candidates. With a small training set of just 288 antibacterial organometallic compounds, the algorithm scanned millions of structures, selecting the most active against resistant bacteria. The most promising candidates were tested and showcased almost six times greater antibiotic activity than the training set.
Antibiotics have become a cornerstone of most modern medicine, as many hospital treatments rely on antibiotics as a measure to control infection, says lead author Angelo Frei from the University of Bern in Switzerland. However, growing bacterial resistance to these drugs has become a serious problem. Recently, researchers have recognised the potential of metal-based antimicrobials including ruthenium complexes. Compared with traditional organic carbon-based chemicals, metal compounds are 10 times more likely to be active against bacteria and are not necessarily more toxic to humans, explains Frei. They represent a vast compound class that has remained largely unexplored for its use in medicine, he adds. Ruthenium compounds are also simple to synthesise, making them easier drug candidates to explore.
Frei says that the team first used a combinatorial chemistry approach developed by co-author Wee Han Ang to create a library of 288 ruthenium compounds, which were then tested against methicillin-resistant Staphylococcus aureus (MRSA). We found a substantial amount to be active (9.4%), and used this data to train machine-learning models to predict the activity against MRSA, he adds. After these first steps, researchers built a virtual library of 77 million ruthenium complexes. The algorithm then identified two million potentially active structures. To verify the predictions, the team assembled a smaller sample of 54 structures and tested them in the lab against MRSA. 53.7% of these compounds were active, which represents a 5.7x higher hit rate than the initial screening, comments Frei.
Organometallic compounds often have distinct mechanisms of action compared to traditional organic antibiotics, which could be advantageous to overcome existing resistance mechanisms, explains Concepcin Gimeno, an expert in metallodrugs at the Institute of Chemical Synthesis and Homogeneous Catalysis in Zaragoza, Spain. Ruthenium complexes interesting properties include biocompatibility and a very low toxicity compared to other metal complexes, adds Gimeno. Ruthenium complexes are already being investigated in clinical trials for cancer.
Nils Metzler-Nolte, an expert in bioinorganic chemistry at Ruhr University Bochum, Germany, admires the versatility of the method. Building upon previous work in combinatorial chemistry by the Ang group a simple one-pot reaction gives over 250 compounds with vastly different 3D shapes and properties, he explains. This is quite unmatched when you consider the three-dimensional space mapped out with these compounds. This is an attractive aspect of organometallic complexes compounds with radically new structures and chemical properties [could offer] antibiotics with new and unprecedented modes of action, says Metzler-Nolte.
Although ruthenium is relatively expensive and scarce, the syntheses are only between one and three steps, which is very economical compared with commercially available drugs. Moreover, the cost of drug discovery and development is not dictated by the cost of the synthesis, but rather by the huge cost of clinical trials, Metzler-Nolte points out.
Follow-up studies will involve a series of experimental and computational validations to confirm and refine the predictions, then synthesis, characterisation, biological tests, iterative design and more, says Gimeno. Perhaps most importantly, molecular simulations will help understand the unusual antibiotic mode of action of these metal complexes, as well as any resistance observed. Some studies show a very low even non-existent development of resistance for metal-based compounds, but I think it would be foolish to underestimate bacteria, says Frei. Our first aim is to generate more data and larger libraries to cover more of the periodic table and predict more specific properties, such as the degree of activity and toxicity.
Follow this link:
Machine learning identifies promising antibacterial ruthenium-based drug candidates - Chemistry World
Using Interpretable Machine Learning to Develop Trading Algorithms – DataDrivenInvestor
11 min read
One problem with many powerful machine learning algorithms is their uninterpretable nature. Algorithms such as neural networks and their many varieties take numbers in and spit numbers out while their inner workings, especially for sufficiently large networks, are impossible to understand. Because of this, its difficult to determine exactly what the algorithms have learned. This non-interpretability loses key information about the structure of the data such as variable importance and variable interactions.
However, other machine learning (ML) algorithms dont suffer these drawbacks. For example, decision trees, linear regression, and general linear regression provide interpretable models with still-powerful predictive capabilities (albeit typically less powerful than more complex models). This post will use a handful of technical indicators as input vectors for this type of ML algorithm to predict buy and sell signals determined by asset returns. The trained models will then be analyzed to determine the importance of the input variables, leading to an understanding of the trading decisions.
For simplicity, indicators readily available from FMPs data API will be used. If replicating, other indicators can easily be added to the dataset and integrated into the model to allow more complex trading decisions.
For demonstration, the indicators used as input to the ML models will be those readily available from FMPs API. A list of these indicators is below.
An n-period simple moving average (SMA) is an arithmetic moving average calculated using the n most recent data points.
FMP Endpoint:
https://financialmodelingprep.com/api/v3/technical_indicator/5min/AAPL?type=sma&period=10
The exponential moving average (EMA), is similar to the SMA but smooths the raw data by applying higher weights to more recent data points.
where S is a smoothing factor, typically 2, and V_t is the value of the dataset at the current time.
Continued here:
Using Interpretable Machine Learning to Develop Trading Algorithms - DataDrivenInvestor
Using Machine Learning and AI in Oncology – Targeted Oncology
James Zou, PhD, assistant professor of biomedical data science at Stanford University, discusses machine learning and the different ways oncologists are utilizing it for the management, treatment, and diagnosis of cancer.
Machine learning is being applied in both early- and late-stage disease, and aids clinicians in providing the best treatment plans and options for their patients with cancer. In this video, Zou further discusses some of the specific methods the algorithm is trained to look at.
Transcription:
0:09 | Machine learning and artificial intelligence are seeing a lot of applications in oncology. For example, in diagnosis, often the clinicians are working with different kinds of imaging data could be mammography images or CT scans. Machine learning AI algorithms can be very helpful in helping clinicians to analyze those kinds of images for them to identify or to segment relevant regions.
0:39 | There are different stages where machine learning is being applied. They will go all the way from early stages in diagnosis to later stages in terms of treatment planning and treatment recommendations. [On the] diagnosis side, we are seeing a lot of these computer vision algorithms, which is a type of AI or machine learning models that are trained to really understand and analyze different images. For example, now there are algorithms that are looking at histopathology images and slides, and then try to diagnose and predict patient outcomes based on those histology images.
1:18 | There are also algorithms that are trained to look at mammography images and try to detect tumors, legions from these mammography images as other diagnosis sites and other treatment planning sites. People also develop machine learning models that look at, for example, mutation profiles of patients, right from their somatic mutations, and then try to predict based on these mutation profiles if immunotherapy or some other treatments are likely to be a good treatment for this particular patient.
Read the original:
Using Machine Learning and AI in Oncology - Targeted Oncology
Machine learning methods to protect banks from risks of complex investment products – Tech Xplore
This article has been reviewed according to ScienceX's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:
by KeAi Communications Co.
close
Artificial intelligence (AI) is frequently touted as a silver bullet to solve complex modeling problems. Among its many applications, it has been investigated as a tool to manage risks of complex investment productsso-called derivative contractsin the investment banking area. Despite the multiple positive reports in this area, concerns have been raised about their practical applicability.
In a new study published in The Journal of Finance and Data Science, a team of researchers from Switzerland and the U.S. explored whether reinforcement learning RL agents can be trained to hedge derivative contracts.
"It should come as no surprise that if you train an AI on simulated market data, it will work well on markets that are reflective of the simulation, and the data consumption of many AI systems is outrageous," explains Loris Cannelli, first author of the study and a researcher at IDSIA in Switzerland.
To overcome the lack of training data, researchers tend to assume an accurate market simulator to train their AI agents. However, setting up such a simulator leads to a classical financial engineering problem: choosing a model to simulate from and its calibration, and making the AI-based approach much like the standard Monte Carlo methods in use for decades.
"Such an AI can also be hardly considered model-free: this would apply only if enough market data was available for training, and this is rarely the case in realistic derivative markets," says Cannelli.
The study, a collaboration between IDSIA and investment bank of UBS, was based on so-called Deep Contextual Bandits, which are well-known in RL for their data-efficiency and robustness. Motivated by operational realities of real-world investment firms, it incorporates end of day reporting requirements and is characterized by a significantly lower training data requirement compared to conventional models, and adaptability to the changing markets.
"In practice, it's the availability of data and operational realities, such as requirements to report end-of-day risk figures, that are the main drivers that dictate the real work at the bank, instead of ideal agent training," clarifies senior author Oleg Szehr, whom, prior to his appointment at IDSIA, was a staff member at several investment banks. "One of the strengths of the newly developed model is that it conceptually resembles business operations at an investment firm and thus is applicable from a practical perspective."
Although the new method is simple, rigorous assessment of model performance demonstrated that the new method outperforms benchmark systems in terms of efficiency, adaptability and accuracy under realistic conditions. "As often the case in real life, less is morethe same applies to risk management too," concludes Cannelli.
More information: Loris Cannelli et al, Hedging using reinforcement learning: Contextual k-armed bandit versus Q-learning, The Journal of Finance and Data Science (2023). DOI: 10.1016/j.jfds.2023.100101
Provided by KeAi Communications Co.
Read this article:
Machine learning methods to protect banks from risks of complex investment products - Tech Xplore
Utilizing a novel high-resolution malaria dataset for climate-informed predictions with a deep learning transformer … – Nature.com
Thomson, M. C. et al. Malaria early warnings based on seasonal climate forecasts from multi-model ensembles. Nature 439(7076), 576579 (2006).
Article ADS CAS PubMed Google Scholar
Hashizume, M., Terao, T. & Minakawa, N. The Indian Ocean Dipole and malaria risk in the highlands of western Kenya. Proc. Natl. Acad. Sci. 106(6), 18571862 (2009).
Article ADS CAS PubMed PubMed Central Google Scholar
Haileselassie, W. et al. Burden of malaria, impact of interventions and climate variability in Western Ethiopia: an area with large irrigation-based farming. BMC Public Health 22(1), 111 (2022).
Article Google Scholar
Zhou, G., Minakawa, N., Githeko, A. K. & Yan, G. Association between climate variability and malaria epidemics in the East African highlands. Proc. Natl. Acad. Sci. 101(8), 23752380 (2004).
Article ADS CAS PubMed PubMed Central Google Scholar
MBra, R. K. et al. Impact of climate variability on the transmission risk of malaria in northern Cte dIvoire. PLoS One 13(6), e0182304 (2018).
Article PubMed PubMed Central Google Scholar
Talapko, J., krlec, I., Alebi, T., Juki, M. & Vev, A. Malaria: the past and the present. Microorganisms 7(6), 179 (2019).
Article CAS PubMed PubMed Central Google Scholar
World Health Organization. World Malaria Report 2020 (World Health Organization, 2020).
Book Google Scholar
Ohrt, C. et al. Information systems to support surveillance for malaria elimination. Am. J. Trop. Med. Hyg. 93(1), 145 (2015).
Article PubMed PubMed Central Google Scholar
Kim, Y. et al. Malaria predictions based on seasonal climate forecasts in South Africa: A time series distributed lag nonlinear model. Sci. Rep. 9(1), 110 (2019).
Google Scholar
Santosh, T., Ramesh, D. & Reddy, D. LSTM based prediction of malaria abundances using big data. Comput. Biol. Med. 124, 103859 (2020).
Article PubMed Google Scholar
Mohapatra, P., Tripathi, N. K., Pal, I. & Shrestha, S. Comparative analysis of machine learning classifiers for the prediction of malaria incidence attributed to climatic factors.
Masinde, M. Africa's Malaria epidemic predictor: Application of machine learning on malaria incidence and climate data. Proc. of the 2020 the 4th International Conference on Compute and Data Analysis. 2937 (2020).
Mussumeci, E. & Coelho, F. C. Large-scale multivariate forecasting models for Dengue-LSTM versus random forest regression. Spatial Spatio Temporal Epidemiol. 35, 100372 (2020).
Article Google Scholar
Rajkomar, A., Dean, J. & Kohane, I. Machine learning in medicine. N. Engl. J. Med. 380(14), 13471358 (2019).
Article PubMed Google Scholar
Nkiruka, O., Prasad, R. & Clement, O. Prediction of malaria incidence using climate variability and machine learning. Inf. Med. Unlocked 22, 100508 (2021).
Article Google Scholar
Thomson, M. C., Mason, S. J., Phindela, T. & Connor, S. J. Use of rainfall and sea surface temperature monitoring for malaria early warning in Botswana. Am. J. Trop. Med. Hyg. 73(1), 214221 (2005).
Article PubMed Google Scholar
Behera, S. K. et al. Malaria incidences in South Africa linked to a climate mode in southwestern Indian Ocean. Environ. Dev.. 27, 4757 (2018).
Article Google Scholar
Eikenberry, S. E. & Gumel, A. B. Mathematical modeling of climate change and malaria transmission dynamics: A historical review. J. Math. Biol. 77(4), 857933 (2018).
Article MathSciNet PubMed Google Scholar
Kifle, M. M. et al. Malaria risk stratification and modeling the effect of rainfall on malaria incidence in Eritrea. J. Environ. Public Health 2019, 111 (2019).
Article Google Scholar
Okuneye, K. & Gumel, A. B. Analysis of a temperature-and rainfall-dependent model for malaria transmission dynamics. Math. Biosci. 287, 7292 (2017).
Article MathSciNet PubMed Google Scholar
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, . & Polosukhin, I. Attention is all you need. In: Advances in neural information processing systems. Vol 30. (2017).
Carmichael, I. & Marron, J. S. Data science vs. statistics: Two cultures?. Jpn. J. Stat. Data Sci. 1(1), 117138 (2018).
Article MathSciNet Google Scholar
Abbasimehr, H. & Baghery, F. S. A novel time series clustering method with fine-tuned support vector regression for customer behavior analysis. Expert Syst. Appl. 204, 117584 (2022).
Article Google Scholar
Xu, J. et al. Forecast of dengue cases in 20 Chinese cities based on the deep learning method. Int. J. Environ. Res. Public Health 17(2), 453 (2020).
Article PubMed PubMed Central Google Scholar
Ho, T. S. et al. Comparing machine learning with case-control models to identify confirmed dengue cases. PLoS Negl. Trop. Dis. 14(11), e0008843 (2020).
Article PubMed PubMed Central Google Scholar
Wang, M. et al. A novel model for malaria prediction based on ensemble algorithms. PloS One 14(12), e0226910 (2019).
Article CAS PubMed PubMed Central Google Scholar
Lim, B., Ark, S. ., Loeff, N. & Pfister, T. Temporal fusion transformers for interpretable multi-horizon time series forecasting. Int. J. Forecast. 37(4), 17481764 (2021).
Article Google Scholar
Susan, S. & Kumar, A. The balancing trick: Optimized sampling of imbalanced datasetsa brief survey of the recent state of the art. Eng. Rep. 3(4), e12298 (2021).
Article Google Scholar
Thickstun, J. The Transformer Model in Equations (University of Washington, 2021).
Google Scholar
Bengio, S., Vinyals, O., Jaitly, N. & Shazeer, N. Scheduled sampling for sequence prediction with recurrent neural networks. Advances in Neural Information Processing Systems. 28 (2015).
Mohapatra, P., Tripathi, N. K., Pal, I. & Shrestha, S. Determining suitable machine learning classifier technique for prediction of malaria incidents attributed to climate of Odisha. Int. J. Environ. Health Res. 32(8), 17161732 (2022).
Article CAS PubMed Google Scholar
Jdey, I., Hcini, G. & Ltifi, H. Deep learning and machine learning for Malaria detection: Overview, challenges and future directions. arXiv preprint arXiv:2209.13292. (2022).
Munir, M., Siddiqui, S. A., Chattha, M. A., Dengel, A. & Ahmed, S. Fusead: Unsupervised anomaly detection in streaming sensors data by fusing statistical and deep learning models. Sensors 19(11), 2451 (2019).
Article ADS PubMed PubMed Central Google Scholar
Kim, M. Prediction of COVID-19 confirmed cases after vaccination: Based on statistical and deep learning models. Sci. Med. J. 3(2), 153165 (2021).
CAS Google Scholar
Martineau, P. et al. Predicting malaria outbreaks from sea surface temperature variability up to 9 months ahead in Limpopo, South Africa, using machine learning. Front. Pub. Health 25(10), 962377 (2022).
Article Google Scholar
Adeola, A. M., Botai, J. O., Olwoch, J. M., Rautenbach, H. C., Adisa, O. M., De Jager, C., Botai, C. M. & Aaron, M. Predicting malaria cases using remotely sensed environmental variables in Nkomazi, South Africa. Geospatial Health. 14(1) (2019).
Mbunge, E., Milham, R. C., Sibiya, M. N. & Jr Takavarasha, S. Machine learning techniques for predicting malaria: Unpacking emerging challenges and opportunities for tackling malaria in sub-saharan Africa. Proc. Computer Science On-line Conference 327344. (Springer International Publishing, Cham, 2023).
Nguyen, V. H. et al. Deep learning models for forecasting dengue fever based on climate data in Vietnam. PLoS Neglect. Trop. Dis. 16(6), e0010509 (2022).
Article Google Scholar
Wu, N., Green, B., Ben, X. & O'Banion, S. Deep transformer models for time series forecasting: The influenza prevalence case. arXiv preprint arXiv:2001.08317. (2020).
Zerveas, G., Jayaraman, S., Patel, D., Bhamidipaty, A. & Eickhoff, C. A transformer-based framework for multivariate time series representation learning. Proc. of the 27th ACM SIGKDD conference on knowledge discovery & data mining 21142124 (2021).
Wang, N. & Zhao, X. Time series forecasting based on convolution transformer. IEICE Trans. Inf. Syst. 106(5), 976985 (2023).
Article Google Scholar
Xu, C., Li, J., Feng, B. & Lu, B. A financial time-series prediction model based on multiplex attention and linear transformer structure. Appl. Sci. 13(8), 5175 (2023).
Article CAS Google Scholar
Ahmed, D. M., Hassan, M. M. & Mstafa, R. J. A review on deep sequential models for forecasting time series data. Appl. Comput. Intell. Soft Comput. 3, 2022 (2022).
Google Scholar
Ahmed, S., Nielsen, I. E., Tripathi, A., Siddiqui, S., Rasool, G. & Ramachandran, R. P. Transformers in time-series analysis: A tutorial. arXiv 2022. arXiv preprint arXiv:2205.01138.
Haugsdal, E., Aune, E. & Ruocco, M. Persistence initialization: A novel adaptation of the transformer architecture for time series forecasting. Appl. Intell. 29, 16 (2023).
Google Scholar
Mohammadi Farsani, R. & Pazouki, E. A transformer self-attention model for time series forecasting. J. Electric. Comput. Eng. Innov. (JECEI) 9(1), 1 (2020).
Google Scholar
Kamana, E., Zhao, J. & Bai, D. Predicting the impact of climate change on the re-emergence of malaria cases in China using LSTMSeq2Seq deep learning model: A modelling and prediction analysis study. BMJ Open. 12(3), e053922 (2022).
Article PubMed PubMed Central Google Scholar
Teklehaimanot, H. D., Schwartz, J., Teklehaimanot, A. & Lipsitch, M. Alert threshold algorithms and malaria epidemic detection. Emerg. Infect. Dis. 10(7), 1220 (2004).
Article PubMed PubMed Central Google Scholar
Hartfield, M. & Alizon, S. Introducing the outbreak threshold in epidemiology. PLoS Pathog. 9(6), e1003277 (2013).
Article CAS PubMed PubMed Central Google Scholar
Bingham, N. H. & Fry, J. M. Regression: Linear Models in Statistics (Springer Science & Business Media, 2010).
Book Google Scholar
Das, A., Kong, W., Sen, R. & Zhou, Y. A decoder-only foundation model for time-series forecasting. arXiv preprint arXiv:2310.10688. (2023).
Radford, A. et al. Language models are unsupervised multitask learners. Open AI Blog. 1(8), 9 (2019).
Google Scholar
NOAA Physical sciences laboratory. NCEP/DOE AMIP-II Reanalysis (Reanalysis-2) Data. NOAA physical sciences laboratory. Available from: https://psl.noaa.gov/data/gridded/ data.ncep.reanalysis2.html. Accessed March 2023.
Liu, M., Ren, S., Ma, S., Jiao, J., Chen, Y., Wang, Z. & Song, W. Gated transformer networks for multivariate time series classification. arXiv preprint arXiv:2103.14438. (2021).
Chu J, Cao J, Chen Y. An ensemble deep learning model based on transformers for long sequence time-series forecasting. Proc. International Conference on Neural Computing for Advanced Applications 273286 (Springer Nature, Singapore, 2022).
Liu, C., Yu, S., Yu, M., Wei, B., Li, B., Li, G. & Huang, W. Adaptive smooth L1 loss: A better way to regress scene texts with extreme aspect ratios. Proc. 2021 IEEE Symposium on Computers and Communications (ISCC) 17 (IEEE, 2021).
Original post:
Utilizing a novel high-resolution malaria dataset for climate-informed predictions with a deep learning transformer ... - Nature.com
How AI, including ChatGPT, is Revolutionizing Healthcare in 2024 – TechiExpert.com
In 2023, AI, like ChatGPT, became a big deal in healthcare. The World Health Organization thinks AI can change healthcare. Now, in 2024, Canadian experts are figuring out how AI will be a big part of healthcare.
AI is making healthcare more personal. Roxana Sultan from the Vector Institute in Toronto says soon AI will look at a lot of information about a patient, not just X-rays. It will look at notes from doctors, lab results, medicines and genetics. This helps not only to find out what is wrong but also to make a special plan for each person.
AI is also making clinical trials faster. These trials test new medicines. Sue Paish from DIGITAL says AI can look at billions of data pieces in a second. This means we can find out fast if new medicines are safe and work well.
But, using AI in healthcare needs good data. If the data is not good, the answers from AI wont be good either. So, it is important to use data from trusted places.
AI is also helping folks take care of their health. Some wear special AI devices that help them check how they are doing, especially people with heart issues. In faraway places, AI is used to check wounds through cellphones. The smart AI sends info to doctors and they can help patients without even meeting them.
But, when we use AI in healthcare, we need to be careful. Dr. Theresa Tam says we need rules to keep patient privacy safe. This means using AI in the right way.
In 2024, healthcare will be more personal and better for patients. But we also need to be careful and use AI in a way that is fair as well as safe for everyone.
Read more here:
How AI, including ChatGPT, is Revolutionizing Healthcare in 2024 - TechiExpert.com