Artificial intelligence in dermatology: the ‘unsupervised’ learning

Acharya, Prakash; Mathur, M

doi:10.1111/bjd.18955

Cited by 3 publications

(4 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They have shown promise in both the fields of dermatology and pathology. 26 27 Deep learning methods in particular work well for large, raw datasets. 28 The studies investigated in our analysis used relatively small, labeled datasets, which makes supervised learning the ideal option.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

The State of Machine Learning in Outcomes Prediction of Transsphenoidal Surgery: A Systematic Review

Yang

Smith

et al. 2022

J Neurol Surg B Skull Base

View full text Add to dashboard Cite

The purpose of this analysis is to assess the use of machine learning (ML) algorithms in the prediction of post-operative outcomes, including complications, recurrence, and death in transsphenoidal surgery. Following PRISMA guidelines, we systematically reviewed all papers that used at least one ML algorithm to predict outcomes after transsphenoidal surgery. We searched Scopus, PubMed, and Web of Science databases for studies published prior to May 12th, 2021. We identified 13 studies enrolling 5048 patients. We extracted the general characteristics of each study; the sensitivity, specificity, AUC of the ML models developed as well as the features identified as important by the ML models. We identified 12 studies with 5048 patients that included ML algorithms for adenomas, three with 1807 patients specifically for acromegaly, and five with 2105 patients specifically for Cushing’s disease. Nearly all were single institution studies. The studies used a heterogeneous mix of ML algorithms and features to build predictive models. All papers reported an AUC greater than .7, which indicates clinical utility. ML algorithms have the potential to predict post-operative outcomes of transsphenoidal surgery and can improve patient care. Ensemble algorithms and neural networks were often top performers when compared to other ML algorithms. Biochemical and pre-operative features were most likely to be selected as important by ML models. Inexplicability remains a challenge, but algorithms such as local interpretable model-agnostic explanation or Shapley value can increase explainability of ML algorithms. Our analysis shows that ML algorithms have the potential to greatly assist surgeons in clinical decision making.

show abstract

Section: Discussionmentioning

confidence: 99%

“…In general, unsupervised algorithms are useful for categorizing uncategorized data, analyzing images or visualizing data and work best with large datasets. They have shown promise in both the fields of dermatology and pathology [27,28].…”

Section: Algorithms and Algorithm Selectionmentioning

confidence: 99%

The State of Machine Learning in Outcomes Prediction of Transsphenoidal Surgery: A Systematic Review

Yang

Smith

et al. 2022

J Neurol Surg B Skull Base

View full text Add to dashboard Cite

show abstract

“…Section 3 has some related surveys associated with the diagnosis of GC.In Section 4, the nature of cancer and its impact on GC, along with operational challenges and limitations. Section 1: Shows the two types of most popular (ML) approaches, as well as their benefits and drawbacks [15]. various types of cancer.…”

Section: Introductionmentioning

confidence: 99%

Diagnosis of Gastric Cancer Using Machine Learning Techniques in Healthcare Sector: A Survey

Jamil

2022

IJCAI

View full text Add to dashboard Cite

Many researchers are trying hard to minimize the incidence of cancers, mainly GC. For GC, the five-year survival rate is generally 5-25%, but for EGC, it is almost up to 90%. Among the cancers, GC is very deadly. It is difficult for doctors to assess its threat to patients as it requires years of medical practice and rigorous testing. The healthcare sector has benefitted from AI for the early diagnosis or classification of GC. However, the current AI-based techniques need further improvement in clinical testing. Heterogeneous GC characterization requires more optimized methods for early detection of GC because of its type and severity. Hence, it is essential to investigate this area further and develop more optimized approaches for early diagnosis. Early detection will increase the chances of successful treatments. In this study, we have conducted a literature survey detailing the role of AI in the healthcare sector for GC diagnosis. We discuss basic principles, advantages and disadvantages, training and testing of data, and integration of applications like DSS, CDSS, KDD, ML, DM, BD, and DL, and their relevance to the healthcare industry. The study focuses on the application of ML techniques used in the diagnosis of GC. This review paper also introduces DM techniques, their application in the healthcare industry, limitations, roles, and operational challenges. These assist pathologists in helping minimize their workload while increasing diagnostic accuracy. These techniques will further assist medical practitioners with their decision-making process. Povzetek: Raziskava o uporabi tehnik strojnega učenja pri diagnostiki raka želodca v zdravstvenem sektorju.

show abstract

“…As the leading technology in the future, the artificial intelligence (AI) is the core driving force of the fourth scientific and technological revolution, which simulates the human thinking process and intelligent behavior to replace or assistant people to complete the works [1,2]. Hardware optimizing design of neural chips is a significant direction to achieve strong AI, and the core component of neural chips is memristor [3].…”

Section: Introductionmentioning

confidence: 99%

Low power consumption photoelectric coupling perovskite memristor with adjustable threshold voltage

Wang¹,

Zhang²,

Xiong³

et al. 2021

Nanotechnology

View full text Add to dashboard Cite

Organic–inorganic halide perovskites (OHPs) have been proven to possess unique optical and electrical properties, and achieved more extensive application as excellent materials for memristors in recent years. Based on the traditional OHP-based memristors, the intermediate layer of the memristor was prepared using yttrium oxide (Y2O3)/OHP stacking structure in this manuscript. The potential barrier between Y2O3 and perovskite is relatively high (ΔE C = 2.13 eV) which leads to comparatively low current of the memristor, thus the power consumption can be reduced. Besides, by changing the external light conditions, one can realize sharp or slow switch between high resistance state (HRS) and low resistance state (LRS), so as to meet the requirement of multilevel data storage, which indicates its promising application prospect in information storage and biological simulation. In addition, based on characteristics of photoelectric coupling, the Y2O3/OHP memristor can also achieve the advantage of adjustable threshold voltage. The transition of HRS and LRS can be realized by changing the illumination condition at any voltage, which means the set and reset voltage are not fixed, so that the memristor with adjustable threshold voltage can adapt to various working conditions.

show abstract

Artificial intelligence in dermatology: the ‘unsupervised’ learning

Abstract: Linked Article: Du‐Harpur et al. Br J Dermatol 2020; DOI:

Cited by 3 publications

References 4 publications

The State of Machine Learning in Outcomes Prediction of Transsphenoidal Surgery: A Systematic Review

The State of Machine Learning in Outcomes Prediction of Transsphenoidal Surgery: A Systematic Review

Diagnosis of Gastric Cancer Using Machine Learning Techniques in Healthcare Sector: A Survey

Low power consumption photoelectric coupling perovskite memristor with adjustable threshold voltage

Contact Info

Product

Resources

About