1
|
Kleinerman A, Rosenfeld A, Benrimoh D, Fratila R, Armstrong C, Mehltretter J, Shneider E, Yaniv-Rosenfeld A, Karp J, Reynolds CF, Turecki G, Kapelner A. Treatment selection using prototyping in latent-space with application to depression treatment. PLoS One 2021; 16:e0258400. [PMID: 34767577 PMCID: PMC8589171 DOI: 10.1371/journal.pone.0258400] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/09/2021] [Accepted: 09/26/2021] [Indexed: 12/28/2022] Open
Abstract
Machine-assisted treatment selection commonly follows one of two paradigms: a fully personalized paradigm which ignores any possible clustering of patients; or a sub-grouping paradigm which ignores personal differences within the identified groups. While both paradigms have shown promising results, each of them suffers from important limitations. In this article, we propose a novel deep learning-based treatment selection approach that is shown to strike a balance between the two paradigms using latent-space prototyping. Our approach is specifically tailored for domains in which effective prototypes and sub-groups of patients are assumed to exist, but groupings relevant to the training objective are not observable in the non-latent space. In an extensive evaluation, using both synthetic and Major Depressive Disorder (MDD) real-world clinical data describing 4754 MDD patients from clinical trials for depression treatment, we show that our approach favorably compares with state-of-the-art approaches. Specifically, the model produced an 8% absolute and 23% relative improvement over random treatment allocation. This is potentially clinically significant, given the large number of patients with MDD. Therefore, the model can bring about a much desired leap forward in the way depression is treated today.
Collapse
Affiliation(s)
| | | | - David Benrimoh
- McGill University, Montreal, Canada
- Aifred Health, Montreal, Canada
| | | | | | | | | | - Amit Yaniv-Rosenfeld
- Shalvata Mental Health Center, Hod Hasharon, Israel
- Tel-Aviv University, Tel-Aviv, Israel
| | - Jordan Karp
- University of Arizona, Tucson, Arizona, United States of America
| | - Charles F. Reynolds
- University of Pittsburgh, Pittsburgh, Pennsylvania, United States of America
| | | | - Adam Kapelner
- Queens College, New York City, NY, United States of America
| |
Collapse
|
2
|
Abstract
The knowledge gained from data mining is highly dependent on the experience of an expert for further analysis to increase effectiveness and wise decision-making. This mined knowledge requires actionability enhancement before it can be applied to real-world problems. The literature highlights the reasons that emerged the need to incorporate human wisdom in decision-making for complex problems. To solve this problem, a domain called ‘Wisdom Mining’ is recommended, proposing a set of algorithms parallel to the algorithms proposed by the data mining. In wisdom mining, a process to extract wisdom needs to be defined with less influence from an expert. This review proposed improvements to data mining techniques and their applications in the real world and emphasised the need to seek ways to harness wisdom from data. This study covers the diverse definitions and different perspectives of wisdom within philosophy, psychology, management and computer science. This comprehensive literature review served as a foundation for constructing a wise decision framework that aided in identifying the wisdom factors like context, utility, location and time. The inclusion of these wisdom factors in existing data mining algorithms makes the transition from data mining to wisdom mining possible. This research includes the relationship between these two mining process that facilitated further elucidation of the wisdom mining process. Potential research trends in the domain are also seen as a potential endeavour to improve the analysis and use of data.
Collapse
Affiliation(s)
- Salma Khan
- Faculty of Engineering & Information Technology, Foundation University Islamabad, Pakistan
| | - Muhammad Shaheen
- Faculty of Engineering & Information Technology, Foundation University Islamabad, Pakistan
| |
Collapse
|
3
|
Peterson JC, Bourgin DD, Agrawal M, Reichman D, Griffiths TL. Using large-scale experiments and machine learning to discover theories of human decision-making. Science 2021; 372:1209-1214. [DOI: 10.1126/science.abe2629] [Citation(s) in RCA: 31] [Impact Index Per Article: 10.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/10/2020] [Accepted: 05/06/2021] [Indexed: 01/13/2023]
Abstract
Predicting and understanding how people make decisions has been a long-standing goal in many fields, with quantitative models of human decision-making informing research in both the social sciences and engineering. We show how progress toward this goal can be accelerated by using large datasets to power machine-learning algorithms that are constrained to produce interpretable psychological theories. Conducting the largest experiment on risky choice to date and analyzing the results using gradient-based optimization of differentiable decision theories implemented through artificial neural networks, we were able to recapitulate historical discoveries, establish that there is room to improve on existing theories, and discover a new, more accurate model of human decision-making in a form that preserves the insights from centuries of research.
Collapse
Affiliation(s)
- Joshua C. Peterson
- Department of Computer Science, Princeton University, Princeton, NJ 08540, USA
| | - David D. Bourgin
- Department of Computer Science, Princeton University, Princeton, NJ 08540, USA
| | - Mayank Agrawal
- Department of Psychology, Princeton University, Princeton, NJ 08540, USA
- Princeton Neuroscience Institute, Princeton University, Princeton, NJ 08544, USA
| | - Daniel Reichman
- Department of Computer Science, Worcester Polytechnic Institute, Worcester, MA 01609, USA
| | - Thomas L. Griffiths
- Department of Computer Science, Princeton University, Princeton, NJ 08540, USA
- Department of Psychology, Princeton University, Princeton, NJ 08540, USA
| |
Collapse
|
4
|
Rosemarin H, Rosenfeld A, Lapp S, Kraus S. LBA: Online Learning-Based Assignment of Patients to Medical Professionals. SENSORS 2021; 21:s21093021. [PMID: 33923098 PMCID: PMC8123356 DOI: 10.3390/s21093021] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 03/24/2021] [Revised: 04/21/2021] [Accepted: 04/22/2021] [Indexed: 11/16/2022]
Abstract
Central to any medical domain is the challenging patient to medical professional assignment task, aimed at getting the right patient to the right medical professional at the right time. This task is highly complex and involves partially conflicting objectives such as minimizing patient wait-time while providing maximal level of care. To tackle this challenge, medical institutions apply common scheduling heuristics to guide their decisions. These generic heuristics often do not align with the expectations of each specific medical institution. In this article, we propose a novel learning-based online optimization approach we term Learning-Based Assignment (LBA), which provides decision makers with a tailored, data-centered decision support algorithm that facilitates dynamic, institution-specific multi-variate decisions, without altering existing medical workflows. We adapt our generic approach to two medical settings: (1) the assignment of patients to caregivers in an emergency department; and (2) the assignment of medical scans to radiologists. In an extensive empirical evaluation, using real-world data and medical experts' input from two distinctive medical domains, we show that our proposed approach provides a dynamic, robust and configurable data-driven solution which can significantly improve upon existing medical practices.
Collapse
Affiliation(s)
- Hanan Rosemarin
- Department of Computer Science, Bar-Ilan University, Ramat Gan 5290002, Israel; (H.R.); (S.L.); (S.K.)
| | - Ariel Rosenfeld
- Department of Information Science, Bar-Ilan University, Ramat Gan 5290002, Israel
- Correspondence:
| | - Steven Lapp
- Department of Computer Science, Bar-Ilan University, Ramat Gan 5290002, Israel; (H.R.); (S.L.); (S.K.)
| | - Sarit Kraus
- Department of Computer Science, Bar-Ilan University, Ramat Gan 5290002, Israel; (H.R.); (S.L.); (S.K.)
| |
Collapse
|
5
|
Shi L, Wang C, Jia H, Hu X. EPS: Robust Pupil Edge Points Selection with Haar Feature and Morphological Pixel Patterns. INT J PATTERN RECOGN 2021. [DOI: 10.1142/s0218001421560024] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/18/2022]
Abstract
Pupil parameters are the essential foundation for many applications, such as cognitive science and human–machine interaction. Existing approaches are still affected by various challenges. We propose a novel pupil detection pipeline (known as Edge Points Selector “EPS”) which is suitable even for partial occlusion, lighting, and specular reflection. EPS consists of pupil area detection, edge selection, and ellipse fitting. For the first time, we find the suitable Haar-like feature of 2D-pupil and a new pupil edge feature in the local pupil area, and integrate them into the proposed pipeline. EPS was compared with two state-of-art methods on 130[Formula: see text]856 images in this work. Within an error threshold of 5 pixels, our method outperforms the comparison algorithms by 33.8% and 19.4%, respectively, on overage.
Collapse
Affiliation(s)
- Lu Shi
- State and Provincial Joint Engineering, Lab. of Advanced Network, Monitoring and Control, Xi’an Technological University, Xi’an, P. R. China
| | - Changyuan Wang
- State and Provincial Joint Engineering, Lab. of Advanced Network, Monitoring and Control, Xi’an Technological University, Xi’an, P. R. China
| | - Hongbo Jia
- Air Force Medical Center of PLA, Air Force Military Medical University, PLA Beijing, P. R. China
| | - Xiuhua Hu
- State and Provincial Joint Engineering, Lab. of Advanced Network, Monitoring and Control, Xi’an Technological University, Xi’an, P. R. China
| |
Collapse
|
6
|
Gear AS, Prakash K, Singh N, Paruchuri P. PredictRV: A Prediction Based Strategy for Negotiations with Dynamically Changing Reservation Value. LECTURE NOTES IN BUSINESS INFORMATION PROCESSING 2020. [PMCID: PMC7215186 DOI: 10.1007/978-3-030-48641-9_10] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Indexed: 10/28/2022]
|
7
|
Abstract
AbstractReinforcement learning (RL) can be extremely effective in solving complex, real-world problems. However, injecting human knowledge into an RL agent may require extensive effort and expertise on the human designer’s part. To date, human factors are generally not considered in the development and evaluation of possible RL approaches. In this article, we set out to investigate how different methods for injecting human knowledge are applied, in practice, by human designers of varying levels of knowledge and skill. We perform the first empirical evaluation of several methods, including a newly proposed method named State Action Similarity Solutions (SASS) which is based on the notion of similarities in the agent’s state–action space. Through this human study, consisting of 51 human participants, we shed new light on the human factors that play a key role in RL. We find that the classical reward shaping technique seems to be the most natural method for most designers, both expert and non-expert, to speed up RL. However, we further find that our proposed method SASS can be effectively and efficiently combined with reward shaping, and provides a beneficial alternative to using only a single-speedup method with minimal human designer effort overhead.
Collapse
|