(Press-News.org) Machine learning is a powerful tool in computational biology, enabling the analysis of a wide range of biomedical data such as genomic sequences and biological imaging. But when researchers use machine learning in computational biology, understanding model behavior remains crucial for uncovering the underlying biological mechanisms in health and disease.
In a recent article in Nature Methods, researchers at Carnegie Mellon University's School of Computer Science propose guidelines that outline pitfalls and opportunities for using interpretable machine learning methods to tackle computational biology problems. The Perspectives article, "Applying Interpretable Machine Learning in Computational Biology — Pitfalls, Recommendations and Opportunities for New Developments," is featured in the journal's August special issue on AI.
"Interpretable machine learning has generated significant excitement as machine learning and artificial intelligence tools are being applied to increasingly important problems," said Ameet Talwalkar, an associate professor in CMU's Machine Learning Department (MLD). "As these models grow in complexity, there is great promise not only in developing highly predictive models but also in creating tools that help end users understand how and why these models make certain predictions. However, it is crucial to acknowledge that interpretable machine learning has yet to deliver turnkey solutions to this interpretability problem."
The paper is a collaboration between doctoral students Valerie Chen in MLD and Muyu (Wendy) Yang in the Ray and Stephanie Lane Computational Biology Department. Chen's earlier work critiquing the interpretable machine learning community's lack of grounding in downstream use cases inspired the article, and the idea was developed through discussions with Yang and Jian Ma, the Ray and Stephanie Lane Professor of Computational Biology.
"Our collaboration began with a deep dive into computational biology papers to survey the application of interpretable machine learning methods," Yang said. "We noticed that many applications used these methods in a somewhat ad hoc manner. Our goal with this paper was to provide guidelines for more robust and consistent use of interpretable machine learning methods in computational biology."
One major pitfall the paper addresses is the reliance on a single interpretable machine learning method. Instead, the researchers recommend using multiple interpretable machine learning methods with diverse sets of hyperparameters and comparing their results to obtain a more comprehensive understanding of the model behavior and its underlying interpretations.
"While some machine learning models seem to work surprisingly well, we often do not fully understand why," Ma said. "In scientific domains like biomedicine, understanding why models work is crucial for discovering fundamental biological mechanisms."
The paper also warns against cherry-picking results when evaluating interpretable machine learning methods, as this can lead to incomplete or biased interpretations of scientific findings.
Chen emphasized that the guidelines may have broader implications for a wider audience of researchers interested in applying interpretable machine-learning methods to their work.
"We hope that machine learning researchers developing new interpretable machine learning methods and tools — particularly those working on explaining large language models — will carefully consider the human-centric aspects of interpretable machine learning," Chen said. "This includes understanding who their target user is and how the method will be used and evaluated."
While understanding model behavior remains crucially important for scientific discovery and a fundamentally unsolved machine learning problem, the authors hope these challenges spur further interdisciplinary collaborations to facilitate the broader use of AI for scientific impact.
END
CMU researchers outline promises, challenges of understanding AI for biological discovery
2024-08-09
ELSE PRESS RELEASES FROM THIS DATE:
Japan SciCom Forum 2024 comes to Fukuoka on October 22-23
2024-08-09
On October 22 and 23, the sixth Japan SciCom Forum Conference (JSF 2024) will be held at Kyushu University's Ito Campus. Fukuoka will become the third city, following Tokyo and Okinawa, to welcome specialists in science communication from Japan and around the world.
JSF 2024 will bring together a diverse group of science communicators, writers, researchers, and journalists, along with experts from overseas. The conference is open to anyone involved in sharing research findings internationally, as well as those interested in science communication, public outreach, and engagement.
This year's JSF will explore a wide range of topics, including ...
Organic farms certified by peers display higher product diversity
2024-08-09
In Brazil, a study compared two systems of organic product certification implemented in São Paulo state. One system involves conventional certification by auditors accredited by the Ministry of Agriculture and the National Institute of Metrology, Quality and Technology (INMETRO). The other is peer-to-peer certification.
The study, reported in an article published in the journal Organic Agriculture, suggests that peer-to-peer certification adds the virtue of agrobiodiversity to organic farming in light of the significantly larger number of products offered by farms with this type of certification. “This ...
Republicans who believe Trump won in 2020 expect significant chaos in November
2024-08-09
Republicans who believe Donald Trump won the 2020 election are anticipating a much more chaotic election cycle this year than other GOP, Democratic, and independent voters, according to new polling data from the SNF Agora Institute at Johns Hopkins University.
Among Republican respondents who believe President Joe Biden did not lawfully win the 2020 election, about 31% think that either “a lot” or “a great deal” of political violence will occur after the 2024 election—compared to 24% of Democratic voters, 21% of independents and just 12% of GOP voters who acknowledge Biden’s victory four years ago, the poll found.
In ...
Memory problems in old age linked to a key enzyme, study in mice finds
2024-08-09
UNIVERSITY PARK, Pa. — Everyone has moments of forgetfulness from time to time, especially as we get older. But older adults don’t just have difficulty remembering new information. They also have a harder time modifying those memories when new details emerge. Yet, little is known about the mechanisms behind memory updating and how those mechanisms go awry with age.
A team of researchers from Penn State has identified an enzyme that contributes to age-related impairments in memory updating. When blocked, older mice were better able to incorporate new information and performed similarly ...
National study shows how internal medicine chief residency has changed over 20 years
2024-08-09
New research shows how the chief resident position in academic internal medicine residency programs has evolved over the past 20 years, revealing how the position has changed, the types of careers these individuals pursue, and improvement in gender representation.
These findings, published this summer by the American Journal of Medicine, stem from a 20-year multicenter study that involved the University of Colorado Internal Medicine Residency Program. CU Department of Medicine faculty member William Turbyfill, MD, was among the study’s site investigators.
Turbyfill, who practices in the Veterans Affairs ...
VA’s Disrupted Care National Project discovers vascular surgery rates still decreasing since COVID-19 pandemic
2024-08-09
White River Junction, VT – Recently published findings from the VA Disrupted Care National Project (DCNP) revealed the number of vascular surgeries performed across the United States continued to decline even after large drops during the COVID-19 pandemic.
A multi-institutional team of researchers, led by the White River Junction VA Medical Center, analyzed 21,031 vascular surgeries of three common procedures from 2019 to 2023 using Medicare claim data. There was a dramatic drop of 47% at the beginning of the pandemic, but while rates of care recovered partially another ...
Looking to boost your heart health? Try a baked potato
2024-08-09
The potato is small enough to fit inside a person’s hand yet contains enough nutrients to whittle waistlines and lower blood sugar in adults with Type 2 diabetes. Yet, despite the fact that potatoes – particularly the skins – are packed with health-boosting nutrients, they routinely get a bad rap among dieters.
That may soon change, thanks to new research by Neda Akhavan, assistant professor in the Department of Kinesiology and Nutrition Sciences within UNLV’s School of Integrated Health ...
Experts provide further proof of role testosterone plays in preventing severe Covid
2024-08-09
A new study has revealed important information about how a patient’s testosterone level can help protect them from severe Covid-19.
Previous research involving Swansea University investigated how sex hormones are likely to be important determinants of Covid-19 severity.
Now digit ratio expert Professor John Manning, of the Applied Sports, Technology, Exercise and Medicine (A-STEM) research team, has been working with colleagues in Poland and Sweden to look more closely at the subject.
He says their findings, which have just been published ...
Vegan diet better than Mediterranean diet for weight loss and reducing harmful inflammatory dietary compounds, finds new research
2024-08-09
WASHINGTON, D.C.—Eating a low-fat vegan diet reduces harmful inflammatory dietary compounds called advanced glycation end-products (AGEs) by 73%, compared to no reduction on a Mediterranean diet, according to new research by the Physicians Committee for Responsible Medicine published in Frontiers in Nutrition. The decrease in AGEs on the vegan diet was associated with an average weight loss of 13 pounds, compared with no change on the Mediterranean diet.
The reduction of dietary AGEs on the low-fat vegan diet came mainly from excluding the consumption ...
'PTNM' System provides new classification for Peyronie's disease and penile curvature
2024-08-09
August 9, 2024 — Analysis of men seen at a specialist clinic provides new insights into the classification of Peyronie's disease (PD) subtypes, according to a report in the September issue of The Journal of Urology®, an Official Journal of the American Urological Association (AUA). The journal is published in the Lippincott portfolio by Wolters Kluwer.
"By analyzing real-world clinical data in a large group of patients, we identified four distinct subtypes of PD, and additional categorization of other causes of penile curvature," comments Landon Trost, MD, of Male Fertility and Peyronie's ...