Page 336«..1020..335336337338..350360..»

More Studies by Columbia Cancer Researchers Are Retracted – The New York Times

Scientists in a prominent cancer lab at Columbia University have now had four studies retracted and a stern note added to a fifth accusing it of severe abuse of the scientific publishing system, the latest fallout from research misconduct allegations recently leveled against several leading cancer scientists.

A scientific sleuth in Britain last year uncovered discrepancies in data published by the Columbia lab, including the reuse of photos and other images across different papers. The New York Times reported last month that a medical journal in 2022 had quietly taken down a stomach cancer study by the researchers after an internal inquiry by the journal found ethics violations.

Despite that studys removal, the researchers Dr. Sam Yoon, chief of a cancer surgery division at Columbia Universitys medical center, and Changhwan Yoon, a more junior biologist there continued publishing studies with suspicious data. Since 2008, the two scientists have collaborated with other researchers on 26 articles that the sleuth, Sholto David, publicly flagged for misrepresenting experiments results.

One of those articles was retracted last month after The Times asked publishers about the allegations. In recent weeks, medical journals have retracted three additional studies, which described new strategies for treating cancers of the stomach, head and neck. Other labs had cited the articles in roughly 90 papers.

A major scientific publisher also appended a blunt note to the article that it had originally taken down without explanation in 2022. This reuse (and in part, misrepresentation) of data without appropriate attribution represents a severe abuse of the scientific publishing system, it said.

Still, those measures addressed only a small fraction of the labs suspect papers. Experts said the episode illustrated not only the extent of unreliable research by top labs, but also the tendency of scientific publishers to respond slowly, if at all, to significant problems once they are detected. As a result, other labs keep relying on questionable work as they pour federal research money into studies, allowing errors to accumulate in the scientific record.

We are having trouble retrieving the article content.

Please enable JavaScript in your browser settings.

Thank you for your patience while we verify access. If you are in Reader mode please exit andlog intoyour Times account, orsubscribefor all of The Times.

Thank you for your patience while we verify access.

Already a subscriber?Log in.

Want all of The Times?Subscribe.

More here:

More Studies by Columbia Cancer Researchers Are Retracted - The New York Times

Read More..

Must Have Checklist For Data Scientists in 2024 – Analytics Insight

In the dynamic landscape of data science, the role of a data scientist continues to evolve, with new technologies, methodologies, and challenges emerging on a regular basis. To thrive in this ever-changing field, data scientists must possess a diverse set of skills that enable them to navigate complex datasets, extract valuable insights, and drive impactful decisions. As we look ahead to 2024, mastering the following ten essential checklist for data scientists to succeed in their roles and make a significant impact in their organizations.

Data scientists must be proficient in programming languages such as Python, R, and SQL to manipulate data, perform analysis, and build models efficiently. A strong foundation in programming enables data scientists to tackle a wide range of data-related tasks effectively.

A deep understanding of statistical concepts and techniques is crucial for data scientists to derive meaningful insights from data. Proficiency in statistical methods such as hypothesis testing, regression analysis, and Bayesian inference allows data scientists to make informed decisions and draw reliable conclusions.

Effective data visualization skills are essential for communicating insights and findings to stakeholders. Data scientists should be proficient in using tools like Matplotlib, Seaborn, and Tableau to create clear and compelling visualizations that enhance understanding and drive action.

Data scientists must possess expertise in machine learning algorithms and techniques to build predictive models and make accurate forecasts. Proficiency in algorithms such as linear regression, decision trees, and neural networks enables data scientists to leverage the power of machine learning for solving complex problems.

A deep understanding of the domain in which they work is critical for data scientists to interpret data effectively and generate actionable insights. Data scientists should possess domain-specific knowledge that allows them to contextualize data, identify relevant patterns, and make informed recommendations.

Data wrangling, or the process of cleaning and transforming raw data into a usable format, is a fundamental skill for data scientists. Proficiency in data wrangling techniques, such as data cleaning, feature engineering, and data preprocessing, ensures that data scientists can work with high-quality data for analysis and modeling.

Data scientists must be adept problem solvers, capable of approaching complex problems with creativity and analytical rigor. Strong problem-solving skills enable data scientists to identify relevant questions, formulate hypotheses, and devise effective strategies for data analysis and interpretation.

Effective collaboration and communication skills are essential for data scientists to work effectively with cross-functional teams and stakeholders. Data scientists should be able to articulate technical concepts in a clear and concise manner, facilitate discussions, and build consensus around data-driven decisions.

In a rapidly evolving field like data science, continuous learning and adaptability are crucial for staying abreast of new technologies and methodologies. Data scientists should embrace a growth mindset, seek out opportunities for learning and development, and adapt to evolving trends and best practices.

Data scientists must adhere to ethical and responsible data practices to ensure the integrity and privacy of data. They should be familiar with regulations such as GDPR and HIPAA, adhere to ethical guidelines for data collection and usage, and prioritize the ethical implications of their work.

In conclusion, mastering these ten skills is essential for data scientists to succeed in their roles and drive impactful outcomes in 2024 and beyond. By continuously developing and honing these skills, data scientists can position themselves as valuable assets to their organizations, making meaningful contributions to the field of data science and driving innovation in their respective domains.

Join our WhatsApp and Telegram Community to Get Regular Top Tech Updates

Link:

Must Have Checklist For Data Scientists in 2024 - Analytics Insight

Read More..

Data Science at Carleton: Researchers Use Big Data to Address Real-World Issues – Carleton Newsroom

Data science is like detective work for information. It uses scientific methods to collect, organize, and analyze data, uncovering useful patterns and insights. As a national hub of data sciences, Carleton University is deeply engaged in data research. With more than 170 data researchers, across five faculties and 26 departments, Carleton is addressing real-world issues including social movements, accessibility, finance, human behavior, and health and wellness.

On March 26, Carleton will host its 10th annual Data Day. Jointly hosted by the Faculty of Science and the Carleton University Institute for Data Science (CUIDS), Data Day is an annual conference that celebrates the latest developments in data science and analytics research.

Read how Carleton data science researchers are making an impact below.

Cryptocurrencies, such as Bitcoin and Ethereum, have garnered significant attention from the media due to their innovative technology and the potential for high returns on investment. However, this media attention can distort market perceptions, create instability, and increase the risk for investors.

Mohamed Al Guindy, a leading researcher in business and financial technology, is spearheading this research to shed light on the relationship between media attention and cryptocurrency price fluctuations.

Mohamed Al Guindy

Its crucial for investors to educate themselves on these issues prior to delving into the unknown. I think this is especially important at a time when there is so much hype, and so little fundamental or useful information, Al Guindy explains.

Al Guindys research is motivated by the extensive media coverage of cryptocurrencies, particularly Bitcoin, and its potential impact on investor behavior. By analyzing tens of millions of tweets, he has found that heightened media coverage and investor attention may contribute to greater price volatility (instability), posing challenges for small investors in the cryptocurrency market.

It might appear easy to follow the crowd and jump on the bandwagon, but that may not be the most optimal choice, Al Guindy says.

Al Guindys work is paving the way for informed decision-making in the cryptocurrency market.

I believe that many of the emerging and most important research questions in business and finance can be addressed using data science and big data.

Read the original:

Data Science at Carleton: Researchers Use Big Data to Address Real-World Issues - Carleton Newsroom

Read More..

3 Under-the-Radar AI Stocks That Could Outshine Nvidia – InvestorPlace

Artificial intelligence (AI) has gained massive popularity over the past year. There is huge interest in AI stocks and every investor is looking for an opportunity to double their money. Nvidia(NASDAQ:NVDA) is a leader in the space and has already made investors rich but many under-the-radar AI stocks can outshine Nvidia.

The global AI market is expected to grow at a 15.83% compound annual growth rate by 2030 and reach amarket volume of $738.80 billion. This means investors have a massive opportunity to benefit from the rising adoption of AI. Nvidia is an expensive stock today, but if you are looking for reasonably priced under-the-radar AI stocks, here are the emerging leaders who could become hot property in the coming months.

Source: Ascannio / Shutterstock.com

At one time, Palantir(NYSE:PLTR) wasnt an investor favorite due to its secrecy. However, the company has gained massive popularity with its AI prowess and artificial intelligence platform (AIP), which is a big hit among companies.

It held boot camps to enable firms to understand how AIP works and there is a massive backlog of boot camps, showing the growing popularity of Palantirs AI product. Up 190% in the year and 48% year-to-date, PLTR stock is trading at $24 and looks undervalued to me.

The U.S. army has chosen Palantir to builda next-generation targeting systemwith a $178 million deal. It already has a strong presence in the defense industry, and Palantirs solid history of exceptional performance makes it an obvious choice for government contracts. However, it has also grown the commercial business by32% year-over-yearin the fourth quarter.

The AI industry is growing, and Palantir could be the next Nvidia. Analysts love the stock and have a positive rating. Wedbush analyst Dan Ives has a price target of $35 and calls it the Lionel Messi of AI. Further,Brian Stutland, a portfolio manager with Equity Armor Investments, considers Palantir a promising AI investment with a price target of $37, a 57% rise from the current level.

Source: JHVEPhoto / Shutterstock.com

Investors have constantly compared Nvidia and Advanced Micro Devices(NASDAQ:AMD). Nvidia has moved ahead in the race, but AMD isnt the one to sit back and watch. The stock has been on a tear since 2023 but is still up 29% YTD and 87% in the year.

Trading at $179, the stock is a solid buy before it starts to soar higher. While you might not see a rally or explosive gains, it will steadily progress. Investors have begun to give up on the stock, but it is too soon. AMD has the potential to bounce back.

The company has launched MI300X, which it believes is the worlds fastest AI hardware, and some of the biggest organizations in the world use its GPUs. Its fourth-quarter results were proof that the company is gaining strength. It saw a 10% YOY rise in sales and a 62% surge in personal computer sales in the quarter.

Several companies in the industry are looking for low-cost alternatives, and this is where AMD chips can win over Nvidia. The PC market is anticipated to improve in the coming months as the demand for AI-enabled PCs will rise. This could work as a catalyst for AMD stock.

KeyBanc analysthas an overweight rating for the stock with a price target of $270, while Melius Researchanalyst has a price target of $265 with a buy rating.

Source: Sundry Photography / Shutterstock.com

ServiceNow(NYSE:NOW) stock is unstoppable, soaring 71% in the year and 11% YTD. Trading at $767, the stock isnt cheap, but it is on its way to becoming the next Nvidia. The company offers solutions to organizations to handle their workforce and customer experience.

It uses AI and workflow management tools that help identify bottlenecks that are impacting the completion of tasks. The company has already signed a partnership with Nvidia to develop AI solutions.

ServiceNow will adopt the latest AI systems of Nvidia to enhance its platform, and Nvidia will expand the services offered by ServiceNow. This means the company will continue to benefit as long as Nvidia keeps growing.

Fundamentally, the company has impressed investors and reported an EPS of $3.11 in the fourth quarter. It beat analyst expectations with a revenue of $2.4 million, up 26% YOY. One of the most important metrics, the subscription revenue, was up 27% in the quarter to hit $2.3 billion, and it closed 168 deals worth $1 million or more in the quarter. This is higher than the 103 deals closed by Palantir.

ServiceNow is trading at a premium, but I believe the first-quarter results will be impressive and could drive the stock higher. There is no stopping the stocks rally.

On the date of publication, Vandita Jadeja did not hold (either directly or indirectly) any positions in the securities mentioned in this article. The opinions expressed in this article are those of the writer, subject to the InvestorPlace.comPublishing Guidelines.

Vandita Jadeja is a CPA and a freelance financial copywriter who loves to read and write about stocks. She believes in buying and holding for long term gains. Her knowledge of words and numbers helps her write clear stock analysis.

Read the original post:

3 Under-the-Radar AI Stocks That Could Outshine Nvidia - InvestorPlace

Read More..

The A.I. Boom Makes Millions for an Unlikely Industry Player: Anguilla – The New York Times

Artificial intelligences integration into everyday life has stirred up doubts and unsettling questions for many about humanitys path forward. But in Anguilla, a tiny Caribbean island to the east of Puerto Rico, the A.I. boom has made the country a fortune.

The British territory collects a fee from every registration for internet addresses that end in .ai, which happens to be the domain name assigned to the island, like .fr for France and .jp for Japan. With companies wanting internet addresses that communicate they are at the forefront of the A.I. boom like Elon Musks X.ai website for his artificial intelligence company Anguilla has recently received a huge influx in requests for domain names.

For each domain registration, Anguillas government gets anywhere from $140 to thousands of dollars from website names sold at auctions, according government data. Last year, Anguillas government made about $32 million from those fees. That amounted to more than 10 percent of gross domestic product for the territory of almost 16,000 people and 35 square miles.

Some people call it a windfall, Anguillas premier, Ellis Webster, said. We just call it God smiling down on us.

Mr. Webster said the government used the money to provide free health care for citizens 70 and older, and it has committed millions of dollars to finish building a school and a vocational training center. The government has also allocated funds to improve its airport; doubled its budget for sports activities, events and facilities; and increased the budget for citizens seeking medical treatment overseas, he said.

The island, which relies heavily on tourism, had been hard hit by the pandemics restrictions on travel and a devastating hurricane in 2017. The .ai domain income was the boost the country needed.

We are having trouble retrieving the article content.

Please enable JavaScript in your browser settings.

Thank you for your patience while we verify access. If you are in Reader mode please exit andlog intoyour Times account, orsubscribefor all of The Times.

Thank you for your patience while we verify access.

Already a subscriber?Log in.

Want all of The Times?Subscribe.

See the rest here:

The A.I. Boom Makes Millions for an Unlikely Industry Player: Anguilla - The New York Times

Read More..

USF plans to launch college focused on artificial intelligence, cybersecurity and computing – University of South Florida

By Adam Freeman, University Communications and Marketing

The University of South Florida has announced its intention to create a college focused on the rapidly evolving fields of artificial intelligence (AI), cybersecurity and computing, with the goal of positioning the Tampa Bay region and state of Florida as a national leader. USF is the first university in Florida and among the first in the nation to announce plans to create a college dedicated to AI and cybersecurity.

The vision for the college would be to offer undergraduate and graduate programs aligned with USFs strategic plan and the states Programs of Strategic Emphasis to prepare students for high-demand careers, empower faculty to conduct innovative research that leads to new discoveries or technological advancements, grow industry partnerships and promote ethical considerations and trust throughout the digital transformation underway in society. Research shows that there has been a five-fold increase in the demand for AI skills with jobs in the U.S., while more than 40% of organizations experiencing a shortage of cybersecurity professionals say they are unable to find enough qualified talent.

The creation of a new college would leverage USF's existing strengths and partnerships in AI, cybersecurity and computing, as well as its location in the Tampa Bay region, a hub for technology and defense industries. At USF, there are approximately 200 faculty members currently engaged in research in related disciplines, which are seeing significant growth in funding awards. Last year the National Science Foundation reportedly awarded more than $800 million for AI-related research.

"As AI and cybersecurity quickly evolve, the demand for professionals skilled in these areas continues to grow, along with the need for more research to better understand how to utilize powerful new technologies in ways that improve our society, USF President Rhea Law said. Through the expertise of our faculty and our strong partnerships with the business community, the University of South Florida is strategically positioned to be a global leader in these fields.

The formation of a new college is subject to continued consultation with faculty through shared governance processes and approval from the USF Board of Trustees. In recent months, an internal task force has been evaluating USFs faculty strengths and exploring opportunities to enhance multidisciplinary collaboration that will help advance USFs academic and research excellence in AI, cybersecurity and computing. The addition of this new college would bring together expertise currently housed across several different colleges to serve as a complement to USFs 13 existing colleges, which would remain operational and continue to be positioned for success.

Establishing this college would align with USFs strategic initiative to enhance academic and research excellence in key areas of societal need and opportunity, said USF Provost and Executive Vice President for Academic Affairs Prasant Mohapatra. By building on our multidisciplinary strengths, such as health, engineering, arts and humanities, and cybersecurity, we aim to support our strategic goals of advancing student success, promoting continuous professional growth, fostering industry, government and community partnerships, and propelling the university towards a top-25 ranking.

More information is available here.

Read more:

USF plans to launch college focused on artificial intelligence, cybersecurity and computing - University of South Florida

Read More..

United Nations adopts U.S.-led resolution to safely develop AI – The Washington Post

The United Nations on Thursday unanimously adopted a resolution to promote the safe and trustworthy development of artificial intelligence, one of the most expansive efforts to date to reach international alignment on the technology.

The United States spearheaded the initiative, spending the last three months negotiating with U.N. member nations. In more than 40 hours of sessions, U.S. officials had lots of heated conversations about the contents of the resolution with U.S. adversaries, including Russia and China, according to a senior Biden administration official, who briefed reporters on the private negotiations on the condition of anonymity.

The United Nations decision to unanimously adopt the resolution underscores how artificial intelligence transcended usual geopolitical divisions, the official said.

We did generally include most of their suggested edits, the official said, referring to Russia, China and Cuba.

The resolution seeks to assert U.S. leadership in the development of AI on the global stage, as the Biden administration increasingly attempts to expand its influence in the intergovernmental body. The European Union and other states are racing ahead of U.S. legislators, who remain in the early stages of crafting AI legislation.

The resolution seeks to promote human rights. However, there are no enforcement mechanisms if countries do not abide by the resolution, and China is already pushing forward with regulations that would require generative AI systems similar to ChatGPT to adhere to the core socialist values.

The broad agreement builds on past international AI agreements. Last year, the United States, China, the European Union, Britain and more than 20 other countries signed onto the so-called Bletchley Declaration, which sought to avoid the existential safety risks of the technology and promote international cooperation on research. However, following criticism that developing nations had been left out of other international AI agreements, the Biden administration pursued a new agreement with the United Nations.

The U.N. resolution comes as nations take divergent paths on regulating AI, following the explosive popularity of ChatGPT and other generative AI tools that can create photos and videos. The European Parliament this month voted to approve Europes AI Act, giving final approval to a law that requires AI developers to disclose data and conduct rigorous testing. President Biden last year signed an AI executive order, but there is no clear path for comprehensive AI legislation in the U.S. Congress.

Read more here:

United Nations adopts U.S.-led resolution to safely develop AI - The Washington Post

Read More..

How Los Angeles is using AI to help people at risk of homelessness – KCRA Sacramento

How Los Angeles is using AI to help people at risk of homelessness

Updated: 5:28 PM PDT Mar 22, 2024

GULSTAN DART. ALL RIGHT TY THANKS FOR THAT. LOS ANGELES COMPANY IS USING ARTIFICIAL INTELLIGENCE TO IDENTIFY WHOS AT RISK OF HOMELESSNESS. AND THEN THEY OFFER RESOURCES TO KEEP THEM IN THEIR HOUSES. THE SOFTWARE TRACKS THINGS LIKE EMERGENCY ROOM VISITS, CRISIS CARE FOR MENTAL HEALTH, SUBSTANCE ABUSE, DISORDER, DIAGNOSIS, ARRESTS AND THEN SIGN UPS FOR PUBLIC BENEFITS LIKE FOOD AID. NOW, OFFICIALS REACH OUT TO THE PERSON THEY OFFER CASH AND SERVICES, AND ONCE APPROVED, THEYRE ASSIGNED A CASEWORKER WHO WORKS WITH THEM FOR 4 TO 6 MONTHS. AND THEYRE ALSO GIVEN 4 TO $6000 TO SPEND ON ESSENTIALS AS WERE A PROGRAM, WHOS USING AI FOR GOOD TO TARGET RESOURCES AT PEOPLE WHO MIGHT OTHERWISE NOT CONNECT TO THOSE RESOURCES, BUT UNDERSTAND THE IMPLICATIONS OF HAVING ACCESS TO SO MUCH DATA AND HOW COMPLICATED THAT SPACE IS, AND JUST HOPE THAT, UM, WE CAN FIND MORE WAYS TO TO SHARE MORE DATA AND USE THAT DATA SHARING TO SUPPORT PROGRAMS THAT REALLY HELP PEOPLE AND TARGET RESOURCES AT THOSE WHO NEED THEM MOST. RIGHT. WELL, THE PROGRAM LAUNCH

How Los Angeles is using AI to help people at risk of homelessness

Updated: 5:28 PM PDT Mar 22, 2024

Los Angeles County is using artificial intelligence to identify who is at risk of homelessness and then offer resources to keep them housed. The software tracks things like emergency room visits, crisis care for mental health, substance abuse disorder diagnosis, arrests and sign-ups for public benefits like food aid. Officials then reach out to offer cash and services. Once approved theyre assigned a case worker who works with them for up to six months. Theyre then given several thousand dollars to spend on essentials. The program launched in 2021 and has helped more than 700 people so far. Dana Vanderford, who leads LA County's new Homelessness Prevention Unit of Housing for Health, talks about the program in the video above. See more coverage of top California stories here | Download our app.

Los Angeles County is using artificial intelligence to identify who is at risk of homelessness and then offer resources to keep them housed.

The software tracks things like emergency room visits, crisis care for mental health, substance abuse disorder diagnosis, arrests and sign-ups for public benefits like food aid.

Officials then reach out to offer cash and services. Once approved theyre assigned a case worker who works with them for up to six months. Theyre then given several thousand dollars to spend on essentials.

The program launched in 2021 and has helped more than 700 people so far.

Dana Vanderford, who leads LA County's new Homelessness Prevention Unit of Housing for Health, talks about the program in the video above.

See more coverage of top California stories here | Download our app.

See the rest here:

How Los Angeles is using AI to help people at risk of homelessness - KCRA Sacramento

Read More..

Nvidia chip, GPT-5, Microsoft AI devices, and Google health redesign top ZDNET’s Innovation Index – ZDNet

Jensen Huang, Nvidia CEO, shows the new Blackwell GPU chip (left).

This week's ZDNET Innovation Index continued to be AI top-heavy with AI-related topics taking three of the top four spots again -- and it was further dominated by some of the largest tech companies in the world and the company that's now synonymous with generative AI.

To recap, the Innovation Index highlights the top trends in tech based on a vote from our panel of journalists and analysts. We're especially looking for the developments that are the most innovative and will have the biggest impact on the future. ZDNET's editorial leaders narrow down the top 10 trends of the week and then our panel votes to rank the top four. If you're not familiar, here's the link to last week's report as well as the inaugural report two weeks ago.

This week's leading trends were:

Our panel's runaway top pick this week was the new Nvidia GPU that's expected to speed up the large language models that power generative AI by as much as 25 times. Speaking of LLMs, the next big one from market leader OpenAI is going to be GPT-5 and the company is already telegraphing that we should expect it to include some new superpowers to automate tasks with AI agents. Not to left out of the AI party, Microsoft took the wraps off its flagship laptop and 2-in-1 tablet that will have a neural processing unit to optimize for AI-related tasks.

And finally, the one non-AI item on this week's list is Google's big redesign of how health conditions are displayed in search results. This is a lot bigger than it sounds since so many people start their health journey on Google when they are having problems or looking to improve something about their health. I have a number of friends who are physicians and they often refer to patients coming in after being told health information of various levels of reliability by "Dr. Google."

Alright, that's it for this week. Check back next week for the latest set of trends.

Visit link:

Nvidia chip, GPT-5, Microsoft AI devices, and Google health redesign top ZDNET's Innovation Index - ZDNet

Read More..

Heterogeneity and predictors of the effects of AI assistance on radiologists – Nature.com

This research complied with all relevant ethical regulations. The study that produced the AI assistance dataset29 used in this study was determined by the Massachusetts Institute of Technology (MIT) Committee on the Use of Humans as Experimental Subjects to be exempt through exempt determination E-2953.

This study used 324 retrospective patient cases from Stanford Universitys healthcare system containing chest X-rays and clinical histories, which include patients indication, vitals and labs. In this study, we analyzed data collected from a total of 140 radiologists participating in two experiment designs. The non-repeated-measure design included 107 radiologists in a non-repeated-measure setup (Supplementary Fig. 1). Each radiologist read 60 patient cases across four subsequences that each contained 15 cases. Each subsequence corresponded to one of four treatment conditions: with AI assistance and clinical histories, with AI assistance and without clinical history, without AI assistance and with clinical histories and without AI assistance and clinical histories. The four subsequences and associated treatment conditions were organized in a random order. The 60 patient cases were randomly selected and randomly assigned to one of the treatment conditions. This design included across-subject and within-subject variations in the treatment conditions; it did not allow within-case-subject comparisons because a case was encountered only once for a radiologist38. Order effects were mitigated by the randomization of treatment conditions. The repeated-measure design included 33 radiologists in a repeated-measure setup (Supplementary Fig. 2). Each radiologist read a total of 60 patient cases, each under each of the four treatment conditions and producing a total of 240 diagnoses. The radiologist completed the experiment in four sessions, and the radiologist read the same 60 randomly selected patient cases in each session under each of the various treatment arms. In each session, 15 cases were read in each treatment arm in batches of five cases. Treatments were randomly ordered. This resulted in the radiologist reading each patient case under a different treatment condition over the four sessions. There was a 2-week washout period15,39,40 between every session to minimize order effects of radiologists reading the same case multiple times. This design included across-subject and within-subject variations as well as across-case-radiologist and within-case-radiologist variations in treatment conditions. Order effects were mitigated by the randomization of treatment conditions. No enrichment was applied to the data collection process. We combined data from both experiment designs from the clinical history conditions. Further details about the data collection process are available in a separate study29, which focuses on establishing a Bayesian framework for defining optimal humanAI collaboration and characterizing actual radiologist behavior in incorporating AI assistance. The study was determined exempt by the MIT Committee on the Use of Humans as Experimental Subjects through exempt determination E-2953.

There are 15 pathologies with corresponding AI predictions: abnormal, airspace opacity, atelectasis, bacterial/lobar pneumonia, cardiomediastinal abnormality, cardiomegaly, consolidation, edema, lesion, pleural effusion, pleural other, pneumothorax, rib fracture, shoulder fracture and support device hardware. These pathologies, the interrelations among these pathologies and additional pathologies without AI predictions can be visualized in a hierarchical structure in Supplementary Fig. B.1. Radiologists were asked to familiarize themselves with the hierarchy before starting, had access to the figure throughout the experiment and had to provide predictions for pathologies following this hierarchy. This aimed to maximize clarity on the specific pathologies referenced in the experiment. When radiologists received AI assistance, they were simultaneously presented with the AI predictions for these 15 pathologies along with the patients chest X-ray and, if applicable, their clinical history. The AI predictions were presented in the form of prediction probabilities on a 0100 scale. The AI predictions were generated by the CheXpert model8, which is a DenseNet121 (ref. 41)-based model for chest X-rays that has been shown to perform similarly to board-certified radiologists. The model generated a single prediction for fracture that was used as the AI prediction for both rib fracture and shoulder fracture. Authors of the CheXpert model8 decided on the 14 pathologies (with a single prediction for fracture) based on the prevalence of observations in radiology reports in the CheXpert dataset and clinical relevance, conforming to the Fleischner Societys recommended glossary42 whenever applicable. Among the pathologies, they included Pneumonia (corresponding to bacterial/lobar pneumonia) to indicate the diagnosis of primary infection and No Finding (corresponding to abnormal) to indicate the absence of all pathologies. These pathologies were set in the creation of the CheXpert labeler8, which has been applied to generate labels for reports in the CheXpert dataset and MIMIC-CXR43, which are among the largest chest X-ray datasets publicly available.

The ground truth probabilities for a patient case were determined by averaging the continuous predicted probabilities of five board-certified radiologists from Mount Sinai Hospital with at least 10years of experience and chest radiology as a subspecialty on a 0100 scale. For instance, if the predicted probabilities of the five board-certified radiologists are 91, 92, 92, 100 and 100, respectively, the ground truth probability is 95. The prevalence of the pathologies based on a ground truth probability threshold of 50 of a pathology being present is shown in Supplementary Table 1.

The participating radiologists represent a diverse set of institutions recruited through two means. Their primary affiliations include large, medium and small clinical settings and non-clinical settings. Additionally, some radiologists are affiliated with an academic hospital, whereas others are not. Radiologists in the non-repeated-measure design were recruited from teleradiology companies. Radiologists in the repeated-measure design were recruited from the Vinmec health system in Vietnam. Details about the participating radiologists and recruitment process can be found in Supplementary Note | Participant recruitment and affiliation.

The experiment interface and instructions presented to participating radiologists can be found in Supplementary Note | Experiment interface and instructions. Before entering the experiment, radiologists were instructed to walk through the experiment instructions, the hierarchy of pathological findings, basic information and performance of the AI model, video demonstration of the experiment interface and examples, consent clauses, comprehension check questions, information on bonus payment that incentivizes effort and practice patient cases covering four treatment conditions and showing example AI predictions from the AI model used in the experiment.

Sex and gender statistics of the participating radiologists and patient cases are available in Supplementary Tables 39 and 40, respectively. Sex and gender were not considered in the original data collection procedures. Disaggregated information about sex and gender at the individual level was collected in the separate study and will be made available29.

We used the empirical Bayes method30 to shrink the raw mean heterogeneous treatment effects and performance metrics of individual radiologists measured on the dataset toward the grand mean to ameliorate overestimating heterogeneity due to sampling error. The values include AIs treatment effects on error, sensitivity and specificity and performance metrics on unassisted error, sensitivity and specificity.

Assume that ({t}_{r}) is radiologist rs true mean treatment effect from AI assistance or any metric of interest. We observe

$$tilde{t}_{r}={t}_{r}+{{{eta }}}_{r}$$

(1)

which differs from ({t}_{r}) by ({{{eta }}}_{r}). We use a normal distribution as the prior distribution over the metric of interest. The mean of the prior distribution can be computed as

$$Eleft[tilde{t}_{r}right]=Eleft[{t}_{r}right],$$

(2)

the mean of the observed mean metric of interest of radiologists. The variance of the prior distribution can be computed as

$$Eleft[{Big({t}_{r}-Eleft[{t}_{r}right]Big)}^{2}right]=Eleft[{left(tilde{t}_{r}-Eleft[tilde{t}_{r}right]right)}^{2}right]-Eleft[{{{eta }}}_{r}^{2}right],$$

(3)

the variance of the observed mean metric of interest of radiologists minus the estimated (Eleft[{{{eta }}}_{r}^{2}right]). We can estimate (Eleft[{{{eta }}}_{r}^{2}right]) with

$$Eleft[{{{eta }}}_{r}^{2}right]=Eleft[{left(frac{1}{{N}_{r}}mathop{sum }limits_{i}{t}_{{ir}}-Eleft[{t}_{{ir}}right]right)}^{2}right]=Eleft[frac{{sum }_{i}{left({t}_{{ir}}-Eleft[{t}_{{ir}}right]right)}^{2}}{{N}_{r}}right]=Eleft[s.e.{left(tilde{t}_{r}right)}^{2}right].$$

(4)

Denote the estimated mean and variance of the prior distribution as ({{rm{mu }}}_{0}) and ({{rm{sigma }}}_{0}^{2}). We can compute the mean of the posterior distribution for radiologist (r) as

$$frac{{{rm{sigma }}}_{r}^{2}{{rm{mu }}}_{0}+{{rm{sigma }}}_{0}^{2}{{rm{mu }}}_{r}}{{{rm{sigma }}}_{0}^{2}+{{rm{sigma }}}_{r}^{2}}$$

(5)

where ({{rm{mu }}}_{r}=widetilde{{t}}_{t}) and ({{rm{sigma }}}_{r}=s.e.left(widetilde{{t}}_{r}right)); we can compute the variance of the posterior as

$$frac{{{rm{sigma }}}_{0}^{2}{{rm{sigma }}}_{r}^{2}}{{{rm{sigma }}}_{0}^{2}+{{rm{sigma }}}_{r}^{2}}$$

(6)

where ({{rm{sigma }}}_{r}=s.e.left(widetilde{{t}}_{r}right)). The updated mean of the posterior distribution is the radiologists metric of interest after shrinkage.

For the analysis on treatment effects on absolute error, we focus on high-prevalence pathologies with prevalence greater than 10%, because radiologists baseline performance without AI assistance is generally highly accurate on low-prevalence pathologies, where they correctly predict that a pathology is not present, and, as a result, there is little variation in radiologists errors. This is especially true when computing each individual radiologists treatment effect. When there is zero variance in the performance of a radiologist under a treatment condition, the associated standard error estimate is zero, making it impossible to perform inference on this radiologists treatment effect.

The combined characteristics model was fitted on a training set of half of the radiologists (n=68) to predict treatment effects of the test set of the remaining half (n=68). The treatment effect predictions on the test set were used as the combined characteristics score for splitting the test set radiologists into binary subgroups (based on whether a particular radiologists combined characteristics score was smaller than or equal to the median treatment effect of radiologists computed from all available reads). Then, the same procedure was repeated after flipping the training set and test set radiologists to split the other set of radiologists into binary subgroups. The experience-based characteristics of radiologists in the randomly split training set and test set were balanced: one set contained 27 radiologists with less than or equal to 6years of experience and 41 radiologists with more than 6years of experience, and the other set contained 41 and 27, respectively. One set contained 47 radiologists who did not specialize in thoracic radiology and 21 radiologists who did, and the other set contained 54 and 14 radiologists, respectively. One set contained 32 radiologists without experience with AI tools and 36 radiologists with experience, and the other set contained 31 and 37, respectively.

To compute a radiologists observed mean treatment effect and the corresponding standard errors and the overall treatment effect of AI assistance across subgroups, we built a linear regression model with the following formulation using the statsmodels library: error1+C(treatment). Here, error refers to the absolute error of a radiologist prediction; 1 refers to an intercept term; and treatment refers to a binary indicator of whether the prediction is made with or without AI assistance. This formulation allows us to compute the treatment effect of AI assistance for both non-repeated-measure and repeated-measure data.

For the analyses on experience-based radiologist characteristics and AI error, we computed the treatment effects of subgroups split based on the predictor of interest by building a linear regression model with the following formulation using the statsmodels library: error1+C(subgroup)+C(treatment):C(subgroup). Here, error refers to the absolute error of a radiologist prediction; 1 refers to an intercept term; subgroup refers to an indicator of the subgroup that the radiologist is split into; and treatment refers to a binary indicator of whether the prediction is made with or without AI assistance. This formulation allows us to compute the subgroup-specific treatment effect of AI assistance for both non-repeated-measure data and repeated-measure data.

To account for correlations of observations within patient cases and radiologists, we computed cluster-robust standard errors that are two-way clustered at the patient case and radiologist level for all inferences unless otherwise specified44,45. With the statsmodels librarys ordinary least squares (OLS) class, we used a clustered covariance estimator as the type of robust sandwich estimator and defined two-way groups based on identifiers of the patient cases and radiologists. The approach assumes that regression model errors are independent across clusters defined by the patient cases and radiologists and adjusts for correlations within clusters.

The reversion to the mean effect and the mechanism of split sampling in avoiding reversion to the mean are explained in the following derivation:

Suppose that ({u}_{i,r}^{* }) and ({a}_{i,r}^{* }) are the true unassisted and assisted diagnostic error of radiologist (r) on patient case i. Suppose that we measure ({u}_{i,r}={u}_{i,r}^{* }+{e}_{i,r}^{u}) and ({a}_{i,r}={a}_{i,r}^{* }+{e}_{i,r}^{a}) where ({e}_{i,r}^{u}) and ({e}_{i,r}^{a}) are measurement errors. Assume that the measurement errors are independent of ({u}_{i,r}^{* }) and ({a}_{i,r}^{* }).

To study the relationship between unassisted error and treatment effect, we intend to build the following linear regression model:

$${u}_{r}^{* }-{a}_{r}^{* }={{beta }}{u}_{r}^{* }+{e}_{r}^{* }$$

(7)

where the error is independent of the independent variable, and ({u}_{r}^{* }) and ({a}_{r}^{* }) are the mean unassisted and assisted performance of radiologist (r). Here, the moment condition

$$Eleft[{e}_{i,r}^{* }times {u}_{i,r}^{* }right]=0$$

(8)

is as desired. This univariate regression estimates the true value of ({{beta }}), which is defined as

$$frac{{rm{Cov}}({{rm{u}}}_{{rm{r}}}^{ast }-{{rm{a}}}_{{rm{r}}}^{ast },,{{rm{u}}}_{{rm{r}}}^{ast })}{{rm{Var}}({{rm{u}}}_{{rm{r}}}^{ast })}$$

(9)

However, because we have access only to noisy measurements ({u}_{r}) and ({a}_{r}), consider instead an approach that builds the model

$${u}_{r}-{a}_{r}={{beta }}{u}_{r}+{e}_{r}$$

(10)

and assumes the moment condition

$$Eleft[{e}_{r}times {u}_{r}right]=0.$$

(11)

This linear regression model using noisy measurements instead generates the following estimate of ({{beta }}):

$$frac{{Cov}left({u}_{r}-{a}_{r},{u}_{r}right)}{{Var}left({u}_{r}right)}=frac{{Cov}left({u}_{r}^{* }-{a}_{r}^{* },{u}_{r}^{* }right)+{Var}left({e}_{r}^{u}right)}{{Var}left({u}_{r}^{* }right)+{Var}left({e}_{r}^{u}right)}$$

(12)

which is incorrect because of the additional ({{V}},{{ar}}left({{{e}}}_{{{r}}}^{{{u}}}right)) terms in the numerator and the denominator. The additional term in the denominator represents attenuation bias, which we address in detail in a later subsection. The term in the numerator represents the reversion to the mean issue, which we now discuss in further detail.

As the equation shows, the bias caused by reversion to the mean is positive. This term exists because the moment condition (Eleft[{e}_{r}times {u}_{r}right]=0), equation (11), is not valid at the true value of ({{beta }}) as shown in the following derivation:

$$begin{array}{c}Eleft[left({u}_{r}-{a}_{r}-{{beta }}{u}_{r}right)times {u}_{r}right]=Eleft[left(left(1-{{beta }}right){u}_{r}-{a}_{r}right)times {u}_{r}right]\ begin{array}{c}=Eleft[left(left(1-{{beta }}right)left({u}_{r}^{* }+{e}_{r}^{u}right)-left({a}_{r}^{* }+{e}_{r}^{a}right)right)times {u}_{r}right]\ begin{array}{c}=Eleft[left(left(left(1-{{beta }}right){u}_{r}^{* }-{a}_{r}^{* }right)+left(1-{{beta }}right){e}_{r}^{u}-{e}_{r}^{a}right)times {u}_{r}right]\ begin{array}{c}=Eleft[left({e}_{r}^{* }+left(1-{{beta }}right){e}_{r}^{u}-{e}_{r}^{a}right)times {u}_{r}right]\ begin{array}{c}=left(1-{{beta }}right)Eleft[{e}_{r}^{u}times {u}_{r}right]\ =left(1-{{beta }}right){Var}left({e}_{r}^{u}right)ne 0.end{array}end{array}end{array}end{array}end{array}$$

Split sampling solves this bias by using separate patient cases for computing unassisted error and treatment effect. A simple construction of split sampling is to use a separate case i for computing the treatment effect and using the remaining cases to compute unassisted error. With this construction, we obtain the following estimate of ({{beta }}):

$$frac{{Cov}left({u}_{i,r}-{a}_{i,r},{u}_{ne i,r}right)}{{Var}left({u}_{ne i,r}right)}$$

(13)

where ({u}_{i,r}) is the unassisted performance on case i for radiologist (r), and ({u}_{ne i,r}) is the mean unassisted performance computed on all unassisted cases other than i. If the errors on each case used to compute ({u}_{r}^{* }) and ({a}_{r}^{* }) are independent, the estimate of ({{beta }}) is equal to

$$frac{{Cov}left({u}_{r}^{* }-{a}_{r}^{* },{u}_{r}^{* }right)}{{Var}left({u}_{ne i,r}right)}$$

(14)

The remaining discrepancy in the denominator again represents attenuation bias and is addressed in a later subsection.

To study unassisted error as a predictor of treatment effect, we built a linear regression model with the following formulation using the statsmodels library: treatment effect1+unassisted error. We designed the following split sampling construction to maximize data efficiency when computing the independent and dependent variables in the linear regression.

Let i index a patient case and (r) index a radiologist. Assume that a radiologist reads ({N}_{u}) cases unassisted and ({N}_{a}) cases assisted. Recall that the unassisted and assisted cases are disjoint for the non-repeated-measure data; they overlap exactly for the repeated-measure data.

For the non-repeated-measure design, we adopt the following construction:

$${u}_{i,r}-{a}_{r}={{beta }}{x}_{ne i,r}+{{rm{varepsilon }}}_{{u}_{i,r}}+{{rm{varepsilon }}}_{{a}_{r}}$$

(15)

where ({x}_{ne i,r}=frac{1}{{N}_{u}-1}{sum }_{kne i}{u}_{k,r}) and ({a}_{r}=frac{1}{{N}_{a}}{sum }_{k}{a}_{k,r}). Here, ({x}_{ne i,r}) is the mean unassisted performance computed on all unassisted cases other than i; ({u}_{{i},{r}}) is the unassisted performance on case i for radiologist (r); and ({a}_{r}) is the mean assisted performance on all assisted cases for radiologist (r).

For the repeated-measure design, we adopt the following construction:

$${u}_{i,r}-{a}_{i,r}={{beta }}{x}_{ne i,r}+{{rm{varepsilon }}}_{{u}_{i,r}}+{{rm{varepsilon }}}_{{a}_{i,r}}$$

(16)

where ({x}_{ne i,r}=frac{1}{{N}_{u}-1}{sum }_{kne i}{u}_{k,r}). Here, ({x}_{ne i,r}) is the mean unassisted performance computed on all cases other than i; ({u}_{i,r}) is the unassisted performance on case i for radiologist (r); and ({a}_{i,r}) is the assisted performance on case i for radiologist (r).

To study unassisted error as a predictor of assisted error, we built a linear regression model with the following formulation using the statsmodels library: assisted error1+unassisted error. We designed the following split sampling construction that maximizes data efficiency when computing the independent and dependent variables in the linear regression.

For the non-repeated-measure design, we adopt the following construction:

$${a}_{i,r}={{beta }}{x}_{r}+{{rm{varepsilon }}}_{i,r}$$

(17)

where ({x}_{r}=frac{1}{{N}_{u}}{sum }_{k},{x}_{k,r}). Here, ({x}_{r}) is the mean unassisted performance computed on all unassisted cases, and ({a}_{i,r}) is the assisted performance on case i for radiologist (r).

For the repeated-measure design, we adopt the following construction:

$${a}_{i,r}={{beta }}{x}_{ne i,r}+{{rm{varepsilon }}}_{i,r}$$

(18)

where ({x}_{ne i,r}=frac{1}{{N}_{u}-1}{sum }_{kne i}{u}_{k,r}). Here, ({x}_{ne i,r}) is the mean unassisted performance computed on all unassisted cases other than i and ({a}_{i,r}) is the assisted performance on case i for radiologist (r).

The constructions above again emphasize the necessity for split sampling. Without split sampling, the mean unassisted performance, which is the independent variable of the linear regression, will be correlated with the error terms due to overlapping patient cases, leading to a bias in the regression.

We adjusted for attenuation bias for the split sampling linear regression formulations.

We want to estimate regressions of the form

$${Y}_{r}={{{beta }}}_{0}+{{{beta }}}_{1}Eleft[{x}_{r}right]+{{rm{varepsilon }}}_{r}$$

(19)

where ({Y}_{r}) is an outcome for radiologist (r) and (Eleft[{x}_{r}right]) is radiologist (r)s average unassisted performance. We observe

$$widetilde{{x}}_{r}=frac{1}{{N}_{r}}mathop{sum }limits_{i}{x}_{{ir}}=Eleft[{x}_{r}right]+{{{eta }}}_{r}$$

(20)

where ({{{eta }}}_{r}=frac{1}{{N}_{r}}mathop{sum }limits_{i}{x}_{{ir}}-Eleft[{x}_{r}right]) and (Eleft[{{{eta }}}_{r}{x}_{r}right]=0) and (Eleft[{{{eta }}}_{r}{{rm{varepsilon }}}_{r}right]=0), which are justified by independent and identically distributed (i.i.d.) sampling of cases and split sampling, respectively.

Using observations from the experiment, we estimate the following regression:

$${Y}_{r}={{rm{gamma }}}_{0}+{{rm{gamma }}}_{1}tilde{x}_{r}+{{rm{varepsilon }}}_{r}$$

(21)

Recall that

$$begin{array}{rcl}{{hat{rm{gamma }}}_{1}}{to }^{p}frac{Eleft[left({x}_{r}+{{{eta }}}_{r}-Eleft[{x}_{r}right]right)left({Y}_{r}-Eleft[{Y}_{r}right]right)right]}{Eleft[{left({x}_{r}+{{{eta }}}_{r}-Eleft[{x}_{r}right]right)}^{2}right]} =\ frac{Eleft[left({x}_{r}-Eleft[{x}_{r}right]right)left({Y}_{r}-Eleft[{Y}_{r}right]right)right]}{Eleft[{left({x}_{r}-Eleft[{x}_{r}right]right)}^{2}right]+Eleft[{{{eta }}}_{r}^{2}right]}={{{beta }}}_{1}{rm{lambda }}end{array}$$

(22)

where ({rm{lambda }}=frac{Eleft[{left({x}_{r}-Eleft[{x}_{r}right]right)}^{2}right]}{Eleft[{left({x}_{r}-Eleft[{x}_{r}right]right)}^{2}right]+Eleft[{{{eta }}}_{r}^{2}right]}) and ({{{beta }}}_{1}=frac{Eleft[left({x}_{r}-Eleft[{x}_{r}right]right)left({Y}_{r}-Eleft[{Y}_{r}right]right)right]}{Eleft[{left({x}_{r}-Eleft[{x}_{r}right]right)}^{2}right]}). We can estimate ({rm{lambda }}) using a plug-in estimator for each term in the data: (1)

$$begin{array}{rcl}Eleft[{{{eta }}}_{r}^{2}right]=Eleft[{left(frac{1}{{N}_{r}}mathop{sum }limits_{i}{x}_{{ir}}-Eleft[{x}_{{ir}}right]right)}^{2}right]\=Eleft[frac{{sum }_{i}{left({x}_{{ir}}-Eleft[{x}_{{ir}}right]right)}^{2}}{{N}_{r}}right]=Eleft[s.e.{left(tilde{x}_{r}right)}^{2}right].end{array}$$

(23)

This is the standard error of the mean estimator. (2)

Read the rest here:

Heterogeneity and predictors of the effects of AI assistance on radiologists - Nature.com

Read More..