Page 1,027«..1020..1,0261,0271,0281,029..1,0401,050..»

The 10 Hottest Data Science And Machine Learning Tools Of 2023 … – CRN

Software News Rick Whiting June 16, 2023, 11:49 AM EDT

Data science and machine learning technologies are in big demand as businesses look for ways to analyze big data and automate data-focused processes. Here are 10 startups with leading-edge data science and machine learning technology that have caught our attention (so far) this year.

Tool Time

Data volumes continue to explode with the global datasphere the total amount of data created, captured, replicated and consumed growing at more than 20 percent a year to reach approximately 291 zettabytes in 2027, according to market researcher IDC.

Efforts by businesses and organizations to derive value from all that data is fueling demand for data science tools and technologies for developing data analysis strategies, preparing data for analysis, developing data visualizations and building data models. (Data science is a field of study that uses a scientific approach to extract meaning and insights from data.)

And more of that data is being used to power machine learning projects, which are becoming ubiquitous within enterprise businesses as they build machine learning models and connect them to operational applications and software features such as personalization and natural language interfaces, notes Daniel Treiman, ML engineering lead at ML platform developer Predibase, in a list of ML predictions for 2023.

All this is spurring demand for increasingly sophisticated data science and machine learning tools and platforms. What follows is a look at 10 hot data science and machine learning tools designed to meet those demands.

Some are from industry giants and more established IT vendors while many are from startups focused exclusively on the data science and machine learning sectors. Some of these are new products introduced over the last year while others are new releases of tools and platforms that offer expanded capabilities to meet the latest demands of this rapidly changing space.

Rick Whiting has been with CRN since 2006 and is currently a feature/special projects editor.Whiting manages a number of CRNs signature annual editorial projects including Channel Chiefs, Partner Program Guide, Big Data 100, Emerging Vendors, Tech Innovators and Products of the Year. He also covers the Big Data beat for CRN. He can be reached at rwhiting@thechannelcompany.com.

Go here to read the rest:
The 10 Hottest Data Science And Machine Learning Tools Of 2023 ... - CRN

Read More..

How machine learning and new AI technologies could change the … – jacksonprogress-argus

The world has never been more online. From work meetings, emails, and texts to shopping, paying bills, and bankingthe possibilities are endless. Technological advances save people time and give companies new tools for growth.

But that connectedness comes with a cost. The internet has also never been more rife with criminals looking for vulnerabilities to exploit, hoping to hold companies hostage with ransomware, executing crafty phishing and social engineering attacks, hacking into proprietary information, or capturing private data such as Social Security numbers and addresses.

Does artificial intelligence or machine learning make it easier or harder for companies to guard against such attacks? Can other improvements provide additional defenses against cyberattacks?

The pros and cons of some of the advances have recently burst into the news. Geoffrey Hinton, the "Godfather of AI,"recently left Google to warn of the dangers of the very technology he helped develop. He worries that generative artificial intelligencewhich can produce text, images, video, and audiowill be used for misinformation and someday even eclipse humans' creativity. Others say those fears are hypothetical.

Drata compiled a list of five technological innovations changing how firms monitor and protect sensitive data essential to their digital operations.

Read the original here:
How machine learning and new AI technologies could change the ... - jacksonprogress-argus

Read More..

Deep Learning: AI, Art History, and the Museum | Magazine – MoMA

MK: Whats interesting is that in Anadols work, older and newer technologies come together in a kind of collage. He customizes a GAN, which is not even the most advanced type of generative AI now, and combines it with different types of rendering software and new mapping algorithms. And he does this in order to play with the degree of efficacy of these different algorithms: to explore different degrees of human intervention and control and different degrees of letting go, letting the machine conjure its own interpretations of the archive and of human memory.

The title of Unsupervised is very specific: it refers to a technical term for a type of machine learning that Anadol and his studio use. Most machine learning is supervised learning, where the AI needs to try to classify the information that is before it. (Even that is still difficult: We have autonomous cars, but they still cant distinguish between the moon and a stop sign.) In supervised learning, humans tag images, for example, or bits of information, in order to train a machine learning model. So, I would go through the data set, and I would tag an image of a pen with the word pen.

Unsupervised learning, on the other hand, is where the machine does the tagging itself. Its a whole other kind of black box where the machine is actually deciding not only how to tag something, what kinds of properties something should be classified as possessing, but its also deciding in many ways what is meaningful, what is of value, in terms of information. This is already opening up a kind of agency on the part of the machine that is very different from traditional processes of supervised learning.

Anadol is using unsupervised learning so that the work can actually generate something new based on its learnings, rather than just classify and process. And then the artist is in his studio working with this model, almost like an electronic musician with lots of different dials in front of him, adjusting what kinds of learning takes place, the rate at which the learning takes place, thousands and thousands of parameters around what kind of forms it could generate.

But at the same time, theres a huge gulf between that stage and getting to the point of creating something that looks the way the works in Unsupervised do. Theres so much intervention and, in fact, human collaboration. Because with machine learning, you often might get noise. It doesnt necessarily generate anything that we find meaningful or that we actually could perceive. And so, Anadol is working in concert with this quasi-organic, changing, adaptive systembut also guiding it away from what it might normally optimize, or think, to produce a series of morphologies that is unpredictable but neither simply a chance occurrence, nor fully automated. Theres an interplay between probability and indeterminacy.

And then, layered on top of that, in two of the works, is a diagram, a visualization, of the AIs movements in space, moving throughout that complex map or galaxy it has constructed based on everything its learned, and classified, and clustered according to patterns of affinities. Again, these are affinities that we may never even perceive or think of, building a very complex mapin this case, literally 1,024 dimensions. This is not perceivable by human eyes. But what Anadol is doing is creating a map of movement that we can perceive, either as a network of shifting, connected lines, or as four-dimensional fluid dynamics, which looks like a rushing waterfall. Its almost as if youre watching a dance unfold in real time, but the choreographic score is being overlaid on top of the dance.

Read the original:
Deep Learning: AI, Art History, and the Museum | Magazine - MoMA

Read More..

Top Python AI and Machine Learning Libraries – TechRepublic

Learn about some of the best Python libraries for programming Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL).

A lot of software developers are drawn to Python due to its vast collection of open-source libraries. Lately, there have been a lot of libraries cropping up in the realm of Machine Learning (ML) and Artificial Intelligence (AI). These libraries can be readily employed by programmers of all levels for tasks in data science, image and data manipulation, and much more. This programming tutorial will shed some light on why Python is the preferred language for Machine Learning and AI as well as list some of the best ML and AI libraries to choose from.

Lead developer for Numerical Python and Pyfort, Paul Dubois, once stated that Python is the most powerful language you can still read.. Other qualities that have helped propel Python to its current station is its versatility and flexibility, which allows Python to be used alongside other programming languages when needed, including powerhouses like Java and C#. On top of that, Python can operate on nearly all OS and platforms on the market.

That might explain Pythons enduring popularity among developers, but why are so many of them choosing Python to work with ML and AI libraries? For starters, the sheer number of ML and AI libraries that are available means that developers can count on finding one for whatever problem needs solving. Moreover, being an Object-oriented programming (OOP) language, Python lends itself particularly well to efficient data use and manipulation.

Here are a few more reasons why Python is among the top programming languages for Machine Learning, Deep Learning, and Artificial Intelligence:

Now that we have discussed why Python is one of the top programming languages, the rest of this article will present some of best python libraries for Machine Learning and AI.

SEE: How to become a Machine Learning Engineer cheat sheet

Formerly known as Numeric, NumPy was the brainchild of Jim Hugunin, along with contributions from several other developers. In 2005, NumPy was officially born when Travis Oliphant incorporated features of the competing Numarray into Numeric, with extensive modifications. Today, NumPy is completely open-source and has many contributors. It is also widely regarded as the best Python library for Machine Learning and AI.

NumPy is mostly utilized by data scientists to perform a variety of mathematical operations on large, multi-dimensional arrays and matrices. NumPy arrays require far less storage area than other Python lists, and they are faster and more convenient to use, making it a great option to increase the performance of Machine Learning models without too much work. Another attractive feature is that NumPy has tools for integrating C, C++, and Fortran code.

Some of NumPys other features that make it popular amongst the scientific community include:

NumPy (see above) is so popular that several libraries are based on it, including SciPy. Like its inspiration, SciPy is also a free, and open-source library. SciPy is geared towards large data sets, as well as the performing of scientific and technical computing against those data sets. SciPy also comes with embedded modules for array optimization and linear algebra, just like NumPy. Playing a key role in scientific analysis and engineering, SciPy has grown to become one of the foundational Python libraries.

The allure of SciPy is that it takes all of NumPys functions and turns them into user-friendly, scientific tools. As such, it is often used for image manipulation and provides basic processing features for high-level, non-scientific mathematical functions.

The main features of SciPy include:

TensorFlow is a free and open source library that is available for Python, JavaScript, C++, and Java. This flexibility lends itself to a wide range of applications in many different sectors. Developed by the Google Brain team for internal Google use in research and production, the initial version was released under the Apache License 2.0 in 2015. Google released the updated version of TensorFlow, named TensorFlow 2.0, in September 2019.

Although TensorFlow can be used for a range of tasks, its particularly adept at the training and inference of deep neural networks. Using TensorFlow, developers can create and train ML models on not just computers but also mobile devices and servers by using TensorFlow Lite and TensorFlow Serving. These alternatives offer the same benefits but for mobile platforms and high-performance servers.

Some of the areas in ML and DL where TensorFlow excels are:

This tutorial shed some light on why Python is the preferred language for Machine Learning and AI and listed some of the best ML and AI libraries to choose from, including TensorFlow, SciPy, and NumPy. We will be adding to this list in the coming weeks so be sure to check back often.

SEE: Learn how to build AI powered software

Go here to read the rest:
Top Python AI and Machine Learning Libraries - TechRepublic

Read More..

Student research puzzles out cryptocurrency risk by comparing … – Bryant University

The recent boom in cryptocurrencies has created a universe of new investment possibilities, not just for individual investors but institutional investors, governments, and publicly listed firms as well with hype to match. Yet as they become more popular, cryptocurrencies have seen enormous fluctuations in price over their relatively short lifespans, adding uncertain risks and returns to the bottom line.

Risk translation: How cryptocurrency impacts company risk, beta and returns, a paper published in the Journal of Capital Markets Studies by Jack Field 23 and Bryant Professor of Finance A. Can Inci, Ph.D., looks beyond both the hype and the doom predictions to gain a true analysis of the novel asset. Through careful investigation, the study, based on Fields honors thesis, compares the various virtues of different crypto-based strategies, including using cryptocurrency as part of a treasury portfolio versus as a medium of exchange or a commission-based asset.

Whether from forces of supply and demand, or from complex algorithmic technologies (such as blockchains), or from a mixture of the two, the underlying worth of cryptocurrencies has been an enigma for investors and politicians alike, the article notes.

RELATED STORY: Inci on the risks, and rewards, of investing in cryptocurrencies

The piece, published in May, examines the effect cryptocurrency assets can have on the risk profiles of publicly traded firms. Through a cross-sectional analysis of the daily returns, volatility, betas and Sharpe ratios of the four largest public holders of cryptocurrencies (MicroStrategy Inc., Tesla, Inc., Square Inc., and Marathon Digital Holdings, Inc.) and five of the largest cryptocurrencies by market cap (Bitcoin, Ether, Tether, Ripple, and Dogecoin), the authors measured the risk and return characteristics of holding cryptocurrencies, as well as the motivations behind holding them as an asset class.

Their conclusions demonstrate the difference in return for different crypto-relate strategies, finding that strategies tailored around the utilization of cryptocurrency as part of a treasury portfolio exhibit the most positive effects on common stock risk and returns, while strategies that use cryptocurrencies as a medium of exchange or a commission-based asset yielded relatively poorer outcomes.

They also note the importance of transparency and risk disclosures in firms dealing with cryptocurrencies. Being such a volatile asset class, cryptocurrencies can introduce uncertainty into a companys balance sheet, as the value of said assets can change drastically in short periods of time. It is necessary not only for a firms managers to understand the implications of cryptocurrencies on total asset values but also for shareholders to have the right to know the true risk in owning equity shares of a company, Field and Inci state.

Now a compliance analyst at Manulife Investment Management, Field chose to focus his thesis on an emerging area. Research on cryptocurrency use in corporate finance is an especially untrodden research area, the authors note, and their study is one of the first on cryptocurrency investments in the treasury departments of publicly traded companies.

Field, who graduated magna cum laude in December with a major in Finance and concentration in Economics, was one of more than 40 students who completed honors thesis projects this year, ranging from analyzing how match results affect the equity value of publicly traded soccer teams to studying the effects of single-use and fabric facemasks on the environment.

In addition to his role as co-author, Inci also served as Fields thesis advisor for his project. Fields editorial advisor was Professor of Finance Hakkan Saraoglu, Ph.D.

View post:
Student research puzzles out cryptocurrency risk by comparing ... - Bryant University

Read More..

Investment Opportunity with Latest Cryptocurrency Miners – GlobeNewswire

NEW YORK, June 16, 2023 (GLOBE NEWSWIRE) -- Crypto enthusiasts around the world have found a safe, convenient, and highly profitable investment opportunity since the launch of Bitmanu miners. Three powerful ASIC mining rigs from this blockchain development company have already claimed their stake as the markets most profitable crypto miners.

Equipped with the latest 3nm chips, Bitmanu miners are capable of delivering higher mining speed in spite of their moderate power consumptions. Above all, these miners offer hash rates that are unheard of in this industry. Naturally, users of these mining rigs find it much easier to earn mining rewards without using a lot of power.

Bitmanu Hash Rates

Many Bitmanu customers have mentioned that needed just a month to earn 100% ROI by mining Monero. The profitability of these rigsis the highest ever in the industry. This has skyrocketed the demand for Bitmanu miners amongst seasoned mining experts as well as newbies looking to build a steady income source.

Potential Profits/Month

Even though Bitmanu miners are extremely efficient and powerful, these machines can be used without any mining knowledge and experience. They are delivered pre-configured, and users can start mining just by connecting them to a power socket.

For the first time in the history of this industry, we have designed mining rigs specifically for the common man. Our ultimate goal is to level the playing field and democratize the market, said David Letoski, CMO of Bitmanu.

To find out more about Bitmanu, please visit https://bitmanu.com/

About Bitmanu: Bitmanu stands as a prominent manufacturing company, driven by a team of investors and renowned experts in the cryptocurrency industry. The company's mission is to make the advantages of the latest technological innovations accessible to everyone. Bitmanu proudly presents an impressive lineup of cryptocurrency miners that deliver exceptional returns on investment with remarkable speed.

See the original post:
Investment Opportunity with Latest Cryptocurrency Miners - GlobeNewswire

Read More..

TechScape: The US is clamping down on cryptocurrency is the UK next? – The Guardian

TechScape

Rishi Sunaks techno-moment has come. Unfortunately for him, it might be too late.

Last week, the US Securities and Exchange Commission (SEC) launched a pair of lawsuits against the countrys biggest cryptocurrency exchanges, Binance and Coinbase.

The lawsuit against Binance, which had been previewed in an earlier action by the CFTC, the US commodities regulator, was juicy:

The SEC complaint alleges that [CEO Changpeng Zhao] directed Binance to conceal the access of high-spending US customers to Binance.com. In one piece of evidence included in the lawsuit, the Binance chief compliance officer messaged a colleague saying: We are operating as a fking unlicensed securities exchange in the USA bro. Elsewhere in the lawsuit, Binances CCO is quoted as saying: We do not want [Binance].com to be regulated ever.

The company runs two supposedly separate exchanges: a regulated US one and an anything-goes international one. A substantial chunk of each lawsuit focuses on the allegation that the company was knowingly helping traders who should only have been allowed on the regulated exchange to skip over to the international one. A Binance spokesperson said: While we take the allegations in the SECs complaint seriously, they should not be the subject of an SEC enforcement action, let alone on an expedited basis. They are unjustified.

But its the lawsuit against Coinbase that is sending shivers through Americas cryptocurrency industry:

Since at least 2019, through the Coinbase platform, Coinbase has operated as an unregistered broker an unregistered exchange and an unregistered clearing agency, the SEC said in its complaint. Coinbase has for years defied the regulatory structures and evaded the disclosure requirements that Congress and the SEC have constructed for the protection of the national securities markets and investors.

Paul Grewal, the chief legal officer and general counsel of Coinbase, said: The SECs reliance on an enforcement-only approach in the absence of clear rules for the digital asset industry is hurting Americas economic competitiveness and companies like Coinbase that have a demonstrated commitment to compliance.

The case against Binance is a clear allegation of clear wrongdoing: if you were to run a crypto exchange that you accept cant service American customers and then you were to secretly help American customers to trade on it, you arent going to be too stunned when regulatory action follows.

But the case against Coinbase is more fundamental. It is the SEC arguing that it is illegal to run a cryptocurrency exchange per se. Specifically, that some unknown number of crypto tokens are, in fact, regulated securities (the SEC names 13 in its suit against Coinbase, including Solana, Cardano and Polygon) and that, even if those projects are not illegal in and of themselves, helping people trade in them is.

Its a controversial assessment. During the ICO boom of 2017, the SEC took action against specific crypto projects that veered too close to the sun, and generally won on the merits: selling a token to investors that looks and acts like a unit of stock, while telling them buy this and youll get rich, is quite easy for a financial regulator to take action on.

But it is less clear that a cryptocurrency exchange where users trade tokens that arent illegal could nonetheless function as an illegal stock exchange. Nonetheless, the industry is hedging its bets and looking for an escape hatch. Enter the UK:

California-based Andreessen Horowitz (A16Z) said Britain was on the right path to becoming a leader in crypto regulation. The venture capital firms new office will open later this year and will be dedicated to investing in crypto and tech startups in the UK and Europe.

Chris Dixon, the head of crypto investing at Andreessen Horowitz, wrote in a blogpost: While there is still work to be done, we believe that the UK is on the right path to becoming a leader in crypto regulation.

The UK also has deep pools of talent, world-leading academic institutions, and a strong entrepreneurial culture.

Rishi Sunak said he was thrilled that the firm had chosen the UK, a move he said was testament to our world-class universities and talent and our strong competitive business environment.

Although the A16Z office is technically targeting the crypto and startup ecosystem in the UK, it will functionally be extremely focused on the crypto part of that mix. The companys latest UK investment is modish crypto-AI startup Gensyn, the office is led by crypto-focused investor Sriram Krishnan, and, well, theres this statement by Sunak:

As we cement the UKs place as a science and tech superpower, we must embrace new innovations like Web3, powered by blockchain technology, which will enable start-ups to flourish here and grow the economy.

That success is founded on having the right regulation and guardrails in place to protect consumers and foster innovation. While theres still work to do, Im determined to unlock opportunities for this technology and turn the UK into the worlds Web3 centre.

Its been a long time coming for the prime minister, who first tried to attach himself to cryptos rising star when he was the chancellor. In 2021, he launched a taskforce to explore a Bank of England digital currency, and a year later, he tasked the Royal Mint with creating an NFT, just as the market imploded. (The plans were dropped just under a year later).

The UK had already been benefiting from the regulatory uncertainty in the US before last weeks actions, with crypto founders viewing it as a comfortable middle-ground between the risk of remaining in the US and the upheaval of relocating to a fully low-touch regime like the UAE. But a one-two punch of a gleefully optimistic prime minister in Britain and the long-awaited arrival of a true crackdown in America could be the impetus needed to spark a substantial relocation.

Of course, there is one problem: Sunaks regime is not long for this world. You would need a higher risk appetite than even your typical angel investor to bet on him staying in power past 2024, and Labour is somewhat less enthusiastic about cryptocurrencies. The gamble, from those in the space with whom Ive spoken to, is less that Sunak will be able to pass friendly laws in the 18 months he has left in office, and more that when hes replaced as prime minister, a crypto clampdown will be extremely low on the list of priorities of whoever replaces him.

A programming note

Im heading off on parental leave next week, to see my son through to his first birthday. I wont be fully absent youll hear from me about once a month for the rest of the year but Ill be joined by a rotating cast of guest writers from around the Guardian and beyond, led by my partner in tech, our global technology editor Dan Milmo.

If you want to read the complete version of the newsletter please subscribe to receive TechScape in your inbox every Tuesday.

{{topLeft}}

{{bottomLeft}}

{{topRight}}

{{bottomRight}}

{{.}}

Read more from the original source:
TechScape: The US is clamping down on cryptocurrency is the UK next? - The Guardian

Read More..

With transparent machine learning tool, engineers accelerate … – University of Wisconsin-Madison

Machine learning can be a powerful tool for discovering and designing new polymers, according to new research from the UWMadison College of Engineering. Photo illustration by Xin (Zoe) Zou/UWMadison College of Engineering

Using the power of prediction, University of WisconsinMadison mechanical engineers have quickly discovered several promising high-performance polymers out of a field of 8 million candidates.

The aerospace, automobile and electronics industries use these polymers, known as polyimides, for a wide variety of applications because they have excellent mechanical and thermal properties including strength, stiffness and heat resistance.

Right now, theres a limited number of existing polyimides because the process of designing them is costly and time-consuming.

However, with their data-driven design framework, the UWMadison engineers leverage machine learning predictions and molecular dynamics simulations to dramatically speed up the discovery of new polyimides with even better properties.

Ying Li

The team detailed its approach in a paper published this month in the Chemical Engineering Journal.

Our findings have broad implications for the field of materials science and will inspire further research in the development of advanced data-driven techniques for materials discovery, says Ying Li, an associate professor of mechanical engineering at UWMadison who led the research. Our design strategy is much more efficient compared to the conventional trial-and-error process and can also be applied to the molecular design of other polymeric materials.

Polyimides are produced through a condensation reaction of dianhydride and diamine/diisocyanate molecules. For their study, the engineers first collected open-source data of the chemical structures of all the existing dianhydride and diamine/diisocyanate molecules, then used that data to build a comprehensive library of 8 million hypothetical polyimides.

Its kind of like building something with LEGO blocks, Li says. You have the basic building blocks a whole bunch of different dianhydride and diamine/diisocyanate molecules. And you could try to build all of the possible structures by hand, but that would take forever because the various combinations are enormous.

So, Li and his colleagues used a computer to combine the building blocks together, which allowed them to organize all possible combinations into a huge database.

Database in hand, the team created multiple machine learning models for the thermal and mechanical properties of polyimides based on experimentally reported values. Using a variety of machine learning techniques, the researchers identified chemical substructures that are most important for determining individual properties.

We incorporated techniques that essentially explain how our machine learning model behaves, so our model isnt a black box, Li says. Weve built a transparent box that allows human experts to immediately understand why the machine learning model made a certain decision.

Applying their well-trained machine learning models, the researchers obtained predictions for the properties of the 8 million hypothetical polyimides. Then they screened that whole dataset and identified the three best hypothetical polyimides with combined properties superior to those of existing polyimides.

They also checked their work: The researchers built all-atom models for their top-three candidates and conducted molecular dynamics simulations to calculate a key thermal property.

The molecular dynamics simulations were in good agreement with the predictions from the machine learning models, so that gives us confidence that our predictions are quite reliable, Li says. In addition, the simulations showed that these new polyimides would be easy to synthesize.

As a final validation method, the team made one of the new polyimides and performed experiments that demonstrated the materials excellent heat resistance. Their experimental results showed the new polyimide could withstand a temperature of about 1,022 degrees Fahrenheit before it started to degrade a result that agreed with their machine learning predictions. In contrast, existing polyimides could endure temperatures only in the range of 392 to 572 degrees F. The researchers also created a web-based application that allows users to explore the new high-performing polyimides with interactive visualization.

Additional authors on the Chemical Engineering Journal paper include equal-contributing first authors Jinlong He of UWMadison, Lei Tao of the University of Connecticut, and Nuwayo Eric Munyaneza of Virginia Polytechnic Institute and State University. Vikas Varshney of the Air Force Research Laboratory, Wei Chen of Northwestern University, and Guoliang Liu of Virginia Polytechnic Institute and State University are also authors on the paper.

The research was supported by funding from the Air Force Office of Scientific Research through the Air Forces Young Investigator Research Program, the Air Force Research Laboratory, and the National Science Foundation.

Read the original here:
With transparent machine learning tool, engineers accelerate ... - University of Wisconsin-Madison

Read More..

Machine-learning-based diagnosis of thyroid fine-needle aspiration … – Nature.com

In this study, a combination of RI image data and color Papanicolaou-stained image data improved the accuracy of MLA for diagnosing cancer using thyroid FNAB specimens. The classification results of the MLA using color Papanicolaou-stained images were highly dependent on the size of the nucleus, but those of the MLA using RI images were less dependent on nucleus size and were affected by information around the nuclear membrane. The final algorithm using data from both types of images together distinguished thyroid cell clusters from benign thyroid nodules and PTC with 100% accuracy.

MLA has shown superior diagnostic performance using images of thyroid FNAB specimens when a convolutional neural network (CNN) architecture was adopted, which is effective for image analysis7,8,12,13. Guan et al.13 studied a CNN-based MLA for classifying hematoxylineosin-stained FNAB specimens of benign thyroid nodule and PTC (TBSRTC II, V and VI). A total of 887 fragmented color images were used in this study, which were cropped from 279 images taken using a digital camera attached to a brightfield microscope. The trained algorithm exhibited 97.7% accuracy for distinguishing between 128 test images of benign and malignant nodules. Range et al.8 used MLA to classify Papanicolaou-stained FNAB specimens of broader spectrum thyroid nodules (TBSRTC IIVI). They used 916 color images obtained using a whole slide scanner. The trained MLA distinguished malignant from benign nodules with high accuracy (90.8%), comparable to that of a pathologist. Similarly, a CNN-based MLA performed well in our study, exhibiting high-accuracy patch-level classification (97.3%) and cluster-level classification (99.0%), using only color Papanicolaou-stained images.

However, given that the purpose of FNAB is to determine whether to operate on thyroid nodules, it must not only exhibit high overall accuracy, but also minimize serious misclassification, such as classification of an obvious malignancy as benign or that of an overtly benign nodule as a malignancy. In Guans study, MLA misclassified some cases that a pathologist classified as obviously benign as a malignancy. Similarly, in Ranges study, MLA misclassified some clearly benign nodules as malignant or misclassified a malignant nodule that was indicated for surgery as benign8. These issues are problematic because they can lead to an erroneous treatment plan for patients who would receive proper treatment if they underwent the current standard care. We studied nodules with relatively distinct benign or malignant characteristics (TBSRTC II, V, and VI). Our findings that RI data improved the accuracy of MLA in these nodules have important clinical significance since these indicate a potential reduction in the aforementioned serious misclassification.

Guan et al.13 suggested that the significant misclassifications of MLA for the thyroid FNAB specimens could be related to the nucleus size. In their study, the cells in false-positive cases showed large nuclei with a high mean pixel color information similar to malignant cells, but the pathologist determined that these cells had a typically benign morphology. The authors interpreted that the classification of MLA was based on the size and staining of the nucleus, but not on the shape. Furthermore, in our results, MLA based on color images showed limitations in accurately classifying benign thyroid cells with a large nucleus or malignant thyroid cells with a small nucleus because the size of the nucleus was the main feature required for classification. However, MLA classification based on the RI image was less affected by nucleus size. This suggests that RI images for can compensate for the limitations of MLA using color images for FNAB specimens whose nuclear size is not typical for benign or malignant cells.

Further results from analyses to explain the models suggest that RI-image based MLA uses the structure and shape of the nucleus for classification. In addition to the algorithm being activated mainly for large nuclei in color images, the algorithm was activated not only by large nuclei but also by nuclei with a clear structure in RI images. The certainty of the MLA classification results was proportional to the detail of the information around the nuclear membrane when based on RI images, but not when based on color images. Detailed nuclear structures, such as nuclear membrane irregularity and micronucleoli are important indicators of thyroid cancer diagnosis26. Thus, the accuracy of MLA classification can be improved when such information is incorporated.

Another potential strength of RI images is the integration of information of a wide vertical space. In a thyroid cytology specimen, cells are scattered over a wide vertical space (i.e. multiple z-plains) rather than over a plane. A single layer (z-plain) 2D image cannot address this vertical spread, and information from out-of-focus cells is likely to be lost or distorted. In contrast, in the RI image obtained through ODT, cells located in different Z-plains are in focus simultaneously. In our study, MLA based on color images showed a false positive result for some out-of-focus patches, whereas MLA based on RI image showed a true negative result for the same image patches (data not shown). However, the out-of-focus area is only a part of the color images, and the use of multiple z-plane images did not improve the accuracy of MLA when compared to the use of a single z-plane image in a previous study8. Therefore, it is unclear whether the aforementioned factor significantly affects the accuracy of MLA.

This study has certain limitations. Despite the large number of sample measurements, this study was performed in a single center and could not cover all conditions of specimens that could exist in real clinical environments. ODT provides optimal RI imaging in un-manipulated living cells27, but we obtained RI images from chromatically stained cells. Staining acted as an extrinsic noise or artifact in the RI images, which reduced the accuracy of MLA. Further study is required to determine the effect of staining on the outcomes. Finally, up to 30% of FNABs may have indeterminate cytopathology (TBSRTC III and IV). This study targeted specimen characteristic of benign or malignant thyroid nodules (TBSRTC II, V, and VI), and therefore, the currently trained algorithm cannot be directly applied to TBSRTC III and IV specimens without relevant training.

To investigate the complementary nature of RI images and color images, a 2D MIP image was generated by projecting the 3D RI image along the z-axis, thereby excluding the influence of dimensionality. Previous studies in the field of cell classification have demonstrated improved performance when using 3D RI images compared to 2D images28,29. Although our research did not incorporate 3D images due to the specific research objectives, we plan to expand our investigations in future studies by incorporating 3D RI images and other 3D imaging modalities.

In this study, we demonstrated the efficacy of multiplexing of RI with standard brightfield imaging using a single ODT platform for MLA-based classification of benign and malignant thyroid FNABs. Multiplexed ODT showed promise for the development of a more accurate classification of thyroid FNABs while reducing the inherent uncertainty and error observed in the current diagnostic standards. Thus, an ODT-based MLA may potentially contribute to an improved cost-effective and rapid point-of-care management of thyroid malignancies.

See the rest here:
Machine-learning-based diagnosis of thyroid fine-needle aspiration ... - Nature.com

Read More..

What to Expect From IRS Cryptocurrency Enforcement – Wealth Management

Practitioners should be paying close attention to tax issues impacting the cryptocurrency industry, according to a recent presentation, Navigating the Crypto Winter Preparing for Cryptocurrency Regulation & Enforcement in Uncertain Times, at the 15th Annual NYU Tax Controversy Forum on June 8, 2023, in New York City.

Regulation and enforcement relating to digital assets has taken some time, but with the influx of funding expected in the coming years to the Internal Revenue Service, its expected that the agency will be cracking down on taxpayers who evade reporting and paying taxes on digital asset transactions.

As a primer, cryptocurrency, non-fungible tokens and other similarly recognized digital assets are classified as property for federal tax purposes and are subject to the same general tax principles. Transactions involving a digital asset are generally required to be reported on a tax return. The IRS has issued guidance on the tax treatment of transactions involving digital assets. While the guidance is rather straightforward on what type of transactions are taxable (as capital gains or income), the speakers spend some time focusing on open issues when it comes to crypto losses.

Open Issues

The speakers started by explaining the various economic loss events for digital assets, including the sale or exchange of the digital asset, abandonment, worthlessness and a distressed or bankrupt crypto exchange. The speakers emphasized that merely because an event has occurred that appears to crystalize an economic loss on a digital asset does not mean that the loss is realized, recognized, and otherwise allowable for US tax purposes. The discussion also focused on some of the shortfalls of the IRS, such as whether theyre really able to monitor crypto transactions (spoiler: the IRS probably wont have much luck with decentralized finance [DeFi]transactions) and how the IRS can definitively know a taxpayer transferred assets to someone else and not just to another account they own. Other unique situations also discussed were what happens when a taxpayer loses a key to a digital wallet only to later find it, how likely is the IRS to go after someone who made a reasonable effort to report when there are so many non-reporters out there and how to characterize staking (as ordinary income or capital gain?) Staking is when you lock crypto assets for a set period of time to help support the operation of a blockchain and earn staking rewards for doing sothe panelists compared it to earning interest but explained that its mechanically different.

Enforcement

After laying out the open issues, the conversation shifted to why this is important for practitioners. It was reiterated that the IRS will continue bolstering its enforcement in this space, leading to more audits. The agency has already updated Form 1040 for the 2022 tax year, asking taxpayers to disclose any transactions of digital assets. Recent enforcement efforts include a successful conviction for conspiracy to launder cryptocurrencies and a court order requiring a bank to produce information concerning U.S. taxpayers who might have failed to report crypto transactions.

Follow-Up Steps

One important takeaway from the presentation is to advise clients to track digital assets and all related information by evaluating what theyve bought and sold. Find out if your client is involved with any DeFi and warn clients that the IRS is taking digital asset reporting very seriously and that its critical they self-report gain/loss even if they dont receive a Form 1099 or a transaction report from an exchange. Lastly, advise clients that crypto isnt as anonymous as they might think. The IRS is already engaging third parties to help it follow digital asset transactions using forensic tracing of the blockchain and is working quickly to enhance its other compliance capabilities. While there still arent robust know your customer and anti-money-laundering policies in place in the crypto space to help fight money laundering and tax evasion, it wont be long before the IRS beefs up its auditing and clients end up in the hot seat if theyre not careful.

Original post:
What to Expect From IRS Cryptocurrency Enforcement - Wealth Management

Read More..