Domain Adaptation and Its Cross-Disciplinary Implications: A Source Guide

Frame the question

Domain adaptation [1] seeks to bridge knowledge gaps between source and target domains, a challenge mirrored in genomic research [2] and AI healthcare applications [4]. This guide examines how these fields intersect through data integration strategies, methodological innovations, and knowledge dissemination frameworks. The comparison reveals shared challenges in handling heterogeneous data and adapting models to new contexts.

What the evidence shows

The review [1] categorizes domain adaptation methods into three approaches: sample-based (weighting observations), feature-based (mapping features), and inference-based (parameter estimation). These methods directly relate to AI healthcare applications [4], which use Deep Learning techniques like GANs and LSTM networks to adapt diagnostic models to new data. Similarly, genomic studies [2] employ statistical methods to infer population histories from ancient DNA, akin to domain adaptation's cross-domain generalization goals. Open Access research [3] emphasizes knowledge-sharing mechanisms that could inform domain adaptation's cross-disciplinary applications, particularly in ensuring equitable access to scientific advancements.

Follow the source trail

The review [1] establishes a foundational framework for domain adaptation, which is corroborated by AI healthcare applications [4] demonstrating its practical implementation. Genomic research [2] provides a biological analogy for domain adaptation's cross-domain generalization challenges, while Open Access studies [3] highlight the importance of knowledge dissemination in scientific innovation. These sources collectively illustrate how domain adaptation principles apply across disciplines, from machine learning to genomics and healthcare.

Use these sources well

For an essay on domain adaptation, begin with [1]'s categorization of methods, then connect to [4]'s AI healthcare applications to show practical implementations. Use [2] to draw parallels with genomic data integration challenges and [3] to discuss knowledge-sharing implications. When discussing limitations, reference [1]'s cross-domain generalization bounds and [3]'s concerns about Open Access sustainability. Always cite all four sources explicitly, as they collectively address the topic's interdisciplinary nature.

What to search next

How might domain adaptation principles [1] address challenges in personalized medicine? What are the ethical implications of AI-driven diagnostics [4] in resource-limited settings? How could Open Access frameworks [3] improve genomic data sharing [2]? What are the limitations of current domain adaptation methods in handling high-dimensional data like genomic sequences?

Verbatim source abstracts

[1] A Review of Domain Adaptation without Target Labels — IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019-10-07, doi:10.1109/tpami.2019.2945942

Domain adaptation has become a prominent problem setting in machine learning and related fields. This review asks the question: How can a classifier learn from a source domain and generalize to a target domain? We present a categorization of approaches, divided into, what we refer to as, sample-based, feature-based, and inference-based methods. Sample-based methods focus on weighting individual observations during training based on their importance to the target domain. Feature-based methods revolve around on mapping, projecting, and representing features such that a source classifier performs well on the target domain and inference-based methods incorporate adaptation into the parameter estimation procedure, for instance through constraints on the optimization procedure. Additionally, we review a number of conditions that allow for formulating bounds on the cross-domain generalization error. Our categorization highlights recurring ideas and raises questions important to further research. [1]

[2] Genomic evidence for the Pleistocene and recent population history of Native Americans — Science, 2015-07-22, doi:10.1126/science.aab3884

How and when the Americas were populated remains contentious. Using ancient and modern genome-wide data, we found that the ancestors of all present-day Native Americans, including Athabascans and Amerindians, entered the Americas as a single migration wave from Siberia no earlier than 23 thousand years ago (ka) and after no more than an 8000-year isolation period in Beringia. After their arrival to the Americas, ancestral Native Americans diversified into two basal genetic branches around 13 ka, one that is now dispersed across North and South America and the other restricted to North America. Subsequent gene flow resulted in some Native Americans sharing ancestry with present-day East Asians (including Siberians) and, more distantly, Australo-Melanesians. Putative "Paleoamerican" relict populations, including the historical Mexican Pericúes and South American Fuego-Patagonians, are not directly related to modern Australo-Melanesians as suggested by the Paleoamerican Model. [2]

[3] The academic, economic and societal impacts of Open Access: an evidence-based review — F1000Research, 2016-09-21, doi:10.12688/f1000research.8460.3

Ongoing debates surrounding Open Access to the scholarly literature are multifaceted and complicated by disparate and often polarised viewpoints from engaged stakeholders. At the current stage, Open Access has become such a global issue that it is critical for all involved in scholarly publishing, including policymakers, publishers, research funders, governments, learned societies, librarians, and academic communities, to be well-informed on the history, benefits, and pitfalls of Open Access. In spite of this, there is a general lack of consensus regarding the potential pros and cons of Open Access at multiple levels. This review aims to be a resource for current knowledge on the impacts of Open Access by synthesizing important research in three major areas: academic, economic and societal. While there is clearly much scope for additional research, several key trends are identified, including a broad citation advantage for researchers who publish openly, as well as additional benefits to the non-academic dissemination of their work. The economic impact of Open Access is less well-understood, although it is clear that access to the research literature is key for innovative enterprises, and a range of governmental and non-governmental services. Furthermore, Open Access has the potential to save both publishers and research funders considerable amounts of financial resources, and can provide some economic benefits to traditionally subscription-based journals. The societal impact of Open Access is strong, in particular for advancing citizen science initiatives, and leveling the playing field for researchers in developing countries. Open Access supersedes all potential alternative modes of access to the scholarly literature through enabling unrestricted re-use, and long-term stability independent of financial constraints of traditional publishers that impede knowledge sharing. However, Open Access has the potential to become unsustainable for research communities if high-cost options are allowed to continue to prevail in a widely unregulated scholarly publishing market. Open Access remains only one of the multiple challenges that the scholarly publishing system is currently facing. Yet, it provides one foundation for increasing engagement with researchers regarding ethical standards of publishing and the broader implications of 'Open Research'. [3]

[4] Artificial Intelligence and COVID-19: Deep Learning Approaches for Diagnosis and Treatment — IEEE Access, 2020-01-01, doi:10.1109/access.2020.3001973

COVID-19 outbreak has put the whole world in an unprecedented difficult situation bringing life around the world to a frightening halt and claiming thousands of lives. Due to COVID-19's spread in 212 countries and territories and increasing numbers of infected cases and death tolls mounting to 5,212,172 and 334,915 (as of May 22 2020), it remains a real threat to the public health system. This paper renders a response to combat the virus through Artificial Intelligence (AI). Some Deep Learning (DL) methods have been illustrated to reach this goal, including Generative Adversarial Networks (GANs), Extreme Learning Machine (ELM), and Long/Short Term Memory (LSTM). It delineates an integrated bioinformatics approach in which different aspects of information from a continuum of structured and unstructured data sources are put together to form the user-friendly platforms for physicians and researchers. The main advantage of these AI-based platforms is to accelerate the process of diagnosis and treatment of the COVID-19 disease. The most recent related publications and medical reports were investigated with the purpose of choosing inputs and targets of the network that could facilitate reaching a reliable Artificial Neural Network-based tool for challenges associated with COVID-19. Furthermore, there are some specific inputs for each platform, including various forms of the data, such as clinical data and medical imaging which can improve the performance of the introduced approaches toward the best responses in practical applications. [4]

Source dossiers

Reference cards for every cited source, using only verified record metadata.

[1] A Review of Domain Adaptation without Target Labels

• Authors: Wouter M. Kouw, Marco Loog
• Venue: IEEE Transactions on Pattern Analysis and Machine Intelligence
• Published: 2019-10-07
• DOI: 10.1109/tpami.2019.2945942
• Citation count: 602
• Institutions: University of Copenhagen; Delft University of Technology
• Topics: Domain Adaptation and Few-Shot Learning, Machine Learning and ELM, Machine Learning and Data Classification, Computer science, Domain adaptation, Artificial intelligence, Adaptation (eye), Domain (mathematical analysis), Computer vision, Pattern recognition (psychology)
• License/access: open access — license unspecified (open access)
• Record: https://doi.org/10.1109/tpami.2019.2945942
• Abstract (verbatim): "Domain adaptation has become a prominent problem setting in machine learning and related fields. This review asks the question: How can a classifier learn from a source domain and generalize to a target domain? We present a categorization of approaches, divided into, what we refer to as, sample-based, feature-based, and inference-based methods. Sample-based methods focus on weighting individual observations during training based on their importance to the target domain. Feature-based methods revolve around on mapping, projecting, and representing features such that a source classifier performs well on the target domain and inference-based methods incorporate adaptation into the parameter estimation procedure, for instance through constraints on the optimization procedure. Additionally, we review a number of conditions that allow for formulating bounds on the cross-domain generalization error. Our categorization highlights recurring ideas and raises questions important to further research." [1]

[2] Genomic evidence for the Pleistocene and recent population history of Native Americans

• Authors: Maanasa Raghavan, Matthias Steinrücken, Kelley Harris, Stephan Schiffels, Simon Rasmussen, Michael DeGiorgio, Anders Albrechtsen, Cristina Valdiosera, María C. Ávila‐Arcos, Anna‐Sapfo Malaspinas, Anders Eriksson, Ida Moltke, Mait Metspalu, Julian R. Homburger, Jeff Wall, Omar E. Cornejo, J. Víctor Moreno-Mayar, Thorfinn Sand Korneliussen, Tracey Pierre, Morten Rasmussen, Paula F. Campos, Peter de Barros Damgaard, Morten E. Allentoft, John Lindo, Ene Metspalu, Ricardo Varela, Josefina Mansilla Lory, Celeste Henrickson, Andaine Seguin‐Orlando, Helena Malmström, Thomas Stafford, Suyash Shringarpure, Andrés Moreno‐Estrada, Monika Karmin, Kristiina Tambets, Anders Bergström, Yali Xue, Vera Warmuth, A. D. Friend, Joy Singarayer, Paul J. Valdes, François Balloux, Ilán Leboreiro, José Luis Vera, Héctor Rangel‐Villalobos, Davide Pettener, Donata Luiselli, Loren G. Davis, Évelyne Heyer, Christoph P. E. Zollikofer, Marcia S. Ponce de León, Colin Smith, Vaughan Grimes, Kelly-Anne Pike, Michael Deal, Benjamin T. Fuller, Bernardo Arriaza, Vivien G. Standen, Maria Francisca Luz, François‐Xavier Ricaut, Niède Guidon, L. P. Osipova, Mikhail I. Voevoda, Olga L. Posukh, Oleg Balanovsky, Maria Lavryashina, Yuri Bogunov, Э. К. Хуснутдинова, Marina Gubina, Elena Balanovska, С.А. Федорова, Sergey Litvinov, B. А. Malyarchuk, М. В. Деренко, M. J. Mosher, David Archer, Jerome S. Cybulski, Barbara Petzelt, Joycelynn Mitchell, Rosita Worl, Paul J. Norman, Peter Parham, Brian M. Kemp, Toomas Kivisild, Chris Tyler-Smith, Manjinder S. Sandhu, Michael Crawford, Richard Villems, David Glenn Smith, Michael R. Waters, Ted Goebel, John R. Johnson, Ripan S. Malhi, Mattias Jakobsson, David J. Meltzer, Andrea Manica, Richard Durbin, Carlos D. Bustamante, Yun S. Song, Rasmus Nielsen
• Venue: Science
• Published: 2015-07-22
• DOI: 10.1126/science.aab3884
• Citation count: 573
• Institutions: University of Copenhagen; Natural History Museum Aarhus; University of Massachusetts Amherst; University of California, Berkeley; Wellcome Sanger Institute; Technical University of Denmark; Pennsylvania State University; La Trobe University
• Topics: Forensic and Genetic Research, Race, Genetics, and Society, Yersinia bacterium, plague, ectoparasites research, Pleistocene, Evolutionary biology, Population, Geography, Biology, Genealogy, History
• License/access: open access — license unspecified (open access)
• Record: https://doi.org/10.1126/science.aab3884
• Abstract (verbatim): "How and when the Americas were populated remains contentious. Using ancient and modern genome-wide data, we found that the ancestors of all present-day Native Americans, including Athabascans and Amerindians, entered the Americas as a single migration wave from Siberia no earlier than 23 thousand years ago (ka) and after no more than an 8000-year isolation period in Beringia. After their arrival to the Americas, ancestral Native Americans diversified into two basal genetic branches around 13 ka, one that is now dispersed across North and South America and the other restricted to North America. Subsequent gene flow resulted in some Native Americans sharing ancestry with present-day East Asians (including Siberians) and, more distantly, Australo-Melanesians. Putative "Paleoamerican" relict populations, including the historical Mexican Pericúes and South American Fuego-Patagonians, are not directly related to modern Australo-Melanesians as suggested by the Paleoamerican Model." [2]

[3] The academic, economic and societal impacts of Open Access: an evidence-based review

• Authors: Jonathan Tennant, François Waldner, Damien Jacques, Paola Masuzzo, Lauren Collister, Chris Hartgerink
• Venue: F1000Research
• Published: 2016-09-21
• DOI: 10.12688/f1000research.8460.3
• Citation count: 570
• Institutions: Imperial College London; UCLouvain; Ghent University; VIB-UGent Center for Medical Biotechnology; University of Pittsburgh; Tilburg University
• Topics: scientometrics and bibliometrics research, Research Data Management Practices, Academic Publishing and Open Access, Viewpoints, Publication, Open science, Scope (computer science), Public relations, Political science, Citation
• License/access: cc-by (open access)
• Record: https://doi.org/10.12688/f1000research.8460.3
• Abstract (verbatim): "Ongoing debates surrounding Open Access to the scholarly literature are multifaceted and complicated by disparate and often polarised viewpoints from engaged stakeholders. At the current stage, Open Access has become such a global issue that it is critical for all involved in scholarly publishing, including policymakers, publishers, research funders, governments, learned societies, librarians, and academic communities, to be well-informed on the history, benefits, and pitfalls of Open Access. In spite of this, there is a general lack of consensus regarding the potential pros and cons of Open Access at multiple levels. This review aims to be a resource for current knowledge on the impacts of Open Access by synthesizing important research in three major areas: academic, economic and societal. While there is clearly much scope for additional research, several key trends are identified, including a broad citation advantage for researchers who publish openly, as well as additional benefits to the non-academic dissemination of their work. The economic impact of Open Access is less well-understood, although it is clear that access to the research literature is key for innovative enterprises, and a range of governmental and non-governmental services. Furthermore, Open Access has the potential to save both publishers and research funders considerable amounts of financial resources, and can provide some economic benefits to traditionally subscription-based journals. The societal impact of Open Access is strong, in particular for advancing citizen science initiatives, and leveling the playing field for researchers in developing countries. Open Access supersedes all potential alternative modes of access to the scholarly literature through enabling unrestricted re-use, and long-term stability independent of financial constraints of traditional publishers that impede knowledge sharing. However, Open Access has the potential to become unsustainable for research communities if high-cost options are allowed to continue to prevail in a widely unregulated scholarly publishing market. Open Access remains only one of the multiple challenges that the scholarly publishing system is currently facing. Yet, it provides one foundation for increasing engagement with researchers regarding ethical standards of publishing and the broader implications of 'Open Research'." [3]

[4] Artificial Intelligence and COVID-19: Deep Learning Approaches for Diagnosis and Treatment

• Authors: Mohammad Jamshidi, Ali Lalbakhsh, Jakub Talla, Zdeněk Peroutka, Farimah Hadjilooei, Pedram Lalbakhsh, Morteza Jamshidi, Luigi La Spada, Mirhamed Mirmozafari, Mojgan Dehghani, Asal Sabet, Saeed Roshani, Sobhan Roshani, Nima Bayat-Makou, Bahare Mohamadzade, Zahra Malek, Alireza Jamshidi, Sara Kiani, Hamed Hashemi‐Dezaki, Wahab Mohyuddin
• Venue: IEEE Access
• Published: 2020-01-01
• DOI: 10.1109/access.2020.3001973
• Citation count: 556
• Institutions: Islamic Azad University of Kermanshah; University of West Bohemia in Pilsen; Macquarie University; Tehran University of Medical Sciences; Razi University; Islamic Azad University Kerman; Edinburgh Napier University; University of Wisconsin–Madison
• Topics: COVID-19 diagnosis using AI, Anomaly Detection Techniques and Applications, Machine Learning and ELM, Computer science, Artificial intelligence, Coronavirus disease 2019 (COVID-19), Adversarial system, Deep learning, Artificial neural network, Machine learning
• License/access: cc-by (open access)
• Record: https://doi.org/10.1109/access.2020.3001973
• Abstract (verbatim): "COVID-19 outbreak has put the whole world in an unprecedented difficult situation bringing life around the world to a frightening halt and claiming thousands of lives. Due to COVID-19's spread in 212 countries and territories and increasing numbers of infected cases and death tolls mounting to 5,212,172 and 334,915 (as of May 22 2020), it remains a real threat to the public health system. This paper renders a response to combat the virus through Artificial Intelligence (AI). Some Deep Learning (DL) methods have been illustrated to reach this goal, including Generative Adversarial Networks (GANs), Extreme Learning Machine (ELM), and Long/Short Term Memory (LSTM). It delineates an integrated bioinformatics approach in which different aspects of information from a continuum of structured and unstructured data sources are put together to form the user-friendly platforms for physicians and researchers. The main advantage of these AI-based platforms is to accelerate the process of diagnosis and treatment of the COVID-19 disease. The most recent related publications and medical reports were investigated with the purpose of choosing inputs and targets of the network that could facilitate reaching a reliable Artificial Neural Network-based tool for challenges associated with COVID-19. Furthermore, there are some specific inputs for each platform, including various forms of the data, such as clinical data and medical imaging which can improve the performance of the introduced approaches toward the best responses in practical applications." [4]