[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-125723-en":3,"doc-seo-125723-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},125723,1649267921044,"Ava Thompson","https://us-avatar.wpscdn.com/avatar/1800007509477c92dfb?_k=1782875107921204101",8,"Research & Report","Towards interpretable quantum machine learning via single-photon quantum walks","Variational quantum algorithms provide a route to quantum machine learning by replacing classical neural networks with parametrized quantum circuits, yet they remain difficult to interpret. The work introduces a variational quantization of projective simulation (PS), a reinforcement learning model designed for interpretable artificial intelligence. PS decisions are formulated as random walks over episodic memory graphs, then implemented as single-photon quantum walks in tunable Mach-Zehnder interferometer lattices. Using a transfer-learning example, the approach shows quantum interference can yield capabilities beyond the classical PS model, while also enabling discussion of how interference supports training and tracing decision making.","arXiv :2301 . 13669v2 [ quant-ph] 16 Oct 2023  \nTowards interpretable quantum machine learning via single-photon quantum walks  \nFulvio Flamini, 1, ∗ Marius Krumm, 1, ∗ Lukas J. Fiderer, 1 Thomas Müller,2 and Hans J. Briegel 1  \n1 Universität Innsbruck, Institut für Theoretische Physik, Technikerstraße 21a, A-6020 Innsbruck, Austria 2 Department of Philosophy, University of Konstanz, Universitätsstraße 10, 78464 Konstanz, Germany  \nVariational quantum algorithms represent a promising approach to quantum machine learning where classical neural networks are replaced by parametrized quantum circuits. However, both approaches suffer from a clear limitation, that is a lack of interpretability. Here, we present a variational method to quantize projective simulation (PS), a reinforcement learning model aimed at interpretable artificial intelligence. Decision making in PS is modeled as a random walk on a graph describing the agent’s memory. To implement the quantized model, we consider quantum walks of single photons in a lattice of tunable Mach-Zehnder interferometers trained via variational algorithms. Using an example from transfer learning, we show that the quantized PS model can exploit quantum interference to acquire capabilities beyond those of its classical counterpart. Finally, we discuss the role of quantum interference for training and tracing the decision making process, paving the way for realizations of interpretable quantum learning agents.  \nI. INTRODUCTION  \nClassical machine learning methods are revolutionizing science and technology, with applications ranging from drug discovery [1] to the study of quantum phases of matter [2] . Consequently, the exploitation of quantum effects to enhance machine learning has emerged as a rapidly evolving research domain [3], showcasing early successes such as provable learning advantages [4] and the development of variational quantum circuits tailored to noisy intermediate-scale quantum devices [5] . These circuits can serve as replacements for classical neural networks within traditional algorithms. Nevertheless, amidst the rapid progress made in both classical and quantum machine learning, it is essential to bear in mind the inherent limitations of classical neural networks, which may persist or even be exacerbated in their quantum counterparts.  \nIn particular, the opaque nature of classical neural networks can prove challenging, especially in scenarios necessitating delicate decision-making directly impacting human lives. Similarly, the quest for deeper comprehension of natural phenomena extends beyond a mere oracle for specific problems: Often the process leading to a solution yields insights that are as valuable as the solution itself. While various approaches have been explored in the machine learning literature [6], the most principled approach appears to be the design of transparent models that enable the tracing of decision processes. However, with the inclusion of quantum effects, the prospect of interpreting decision processes appears daunting. Interesting first steps in this direction, using Shapley values to measure feature importance [7] or mutual information to trace the flow of information [8], still rely on black-box methods from classical explainability research.  \nIn this work, we aim to dispel such reservations by quantizing a transparent classical reinforcement learning model known as projective simulation (PS) [9, 10] . By undertaking this quantization, we lay the groundwork for investigating the interpretability aspects inherent in quantum machine learning  \n∗ These authors contributed equally to this work. [marius.krumm@uibk.ac.at](marius.krumm@uibk.ac.at)  \nmodels. In PS, the agent’s decision making process is modeled as a random walk of an excitation in an episodic memory. Following an approach already investigated by some of the authors [11], we replace this classical random walk with a single-photon quantum walk in an optical interferometer. On the one ","cbCaipbrl3mouvxi","https://ap.wps.com/l/cbCaipbrl3mouvxi","pdf",2460886,1,19,"English","en",105,"# Introduction\n# Projective Simulation\n# Quantizing Projective Simulation\n## Photonic Implementation\n## Variational Training Algorithms\n# Numerical Evidence and Generalization\n# Software Package and Extensions","[{\"question\":\"What interpretability problem does the work target in quantum machine learning?\",\"answer\":\"The work addresses the lack of interpretability that affects both variational quantum algorithms and classical neural-network approaches, aiming to enable tracing of decision processes in quantum learning.\"},{\"question\":\"How is projective simulation (PS) decision making modeled?\",\"answer\":\"PS models an agent’s decisions as a random walk of an excitation over a graph representing episodic memory, which defines the policy through graph dynamics.\"},{\"question\":\"How are quantum walks implemented in the proposed quantized PS model?\",\"answer\":\"The quantized PS model is implemented using single-photon quantum walks in a lattice of tunable Mach-Zehnder interferometers, trained with variational algorithms.\"}]","Towards interpretable quantum machine learning via single-photon quantum walks | PDF",1785900858,48,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"towards-interpretable-quantum-machine-learning-via-single-photon-quantum-walks","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/towards-interpretable-quantum-machine-learning-via-single-photon-quantum-walks/125723/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-05",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What interpretability problem does the work target in quantum machine learning?","Question",{"text":75,"@type":76},"The work addresses the lack of interpretability that affects both variational quantum algorithms and classical neural-network approaches, aiming to enable tracing of decision processes in quantum learning.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How is projective simulation (PS) decision making modeled?",{"text":80,"@type":76},"PS models an agent’s decisions as a random walk of an excitation over a graph representing episodic memory, which defines the policy through graph dynamics.",{"name":82,"@type":73,"acceptedAnswer":83},"How are quantum walks implemented in the proposed quantized PS model?",{"text":84,"@type":76},"The quantized PS model is implemented using single-photon quantum walks in a lattice of tunable Mach-Zehnder interferometers, trained with variational algorithms.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]