IA Scholar Query: Complexity of Unification in Free Groups and Free Semi-groups
https://scholar.archive.org/
Internet Archive Scholar query results feedeninfo@archive.orgSat, 31 Dec 2022 00:00:00 GMTfatcat-scholarhttps://scholar.archive.org/help14401 The University and the Arab World
https://scholar.archive.org/work/7nwa6fyht5a3vozk26updrkami
work_7nwa6fyht5a3vozk26updrkamiSat, 31 Dec 2022 00:00:00 GMTEchoes of Rituals of Initiation and Blood Sacrifice in Heart of Darkness by Joseph Conrad
https://scholar.archive.org/work/kugfwnsv7neqdh57x3vt4glifa
In Heart of Darkness, the protagonist Kurtz, of whom we do not, in fact, see much, is shown as connected with a native "sorcerer," a "witch-man," who had "antelope horns" on his head. Antelopes, or goats, are typical sacrificial animals, and the protagonist of this novella is a European who perishes in the midst of tropical forests, in spite of the high hopes that accompany his decision to try his luck in an exotic environment. Kurtz has promising beginnings, but later he gradually degenerates, carrying out what may be called a reversal of the ritual of initiation, comparable to the inverted ritual, to use V. Propp's term, in folklore. In this sense, he may be regarded as a counterpoint to Conrad himself whose life can easily be described as a modern and uncommonly successful enactment of the same ritual. Meanwhile, Kurtz's, and to a lesser extent also Marlow's, failure as initiates is inscribed in the failure of the European civilization to construct a European empire in Africa.Andrzej Wicherwork_kugfwnsv7neqdh57x3vt4glifaThu, 24 Nov 2022 00:00:00 GMTA proposal for 3d quantum gravity and its bulk factorization
https://scholar.archive.org/work/cycwrprrijgplluxitariaiu7e
Recent progress in AdS/CFT has provided a good understanding of how the bulk spacetime is encoded in the entanglement structure of the boundary CFT. However, little is known about how spacetime emerges directly from the bulk quantum theory. We address this question in an effective 3d quantum theory of pure gravity, which describes the high temperature regime of a holographic CFT. This theory can be viewed as a q-deformation and dimensional uplift of JT gravity. Using this model, we show that the Bekenstein-Hawking entropy of a two-sided black hole equals the bulk entanglement entropy of gravitational edge modes. In the conventional Chern-Simons description, these black holes correspond to Wilson lines in representations of (2,ℝ)⊗(2,ℝ). We show that the correct calculation of gravitational entropy suggests we should interpret the bulk theory as an extended topological quantum field theory associated to the quantum semi-group ^+_q(2,ℝ)⊗^+_q(2,ℝ). Our calculation suggests an effective description of bulk microstates in terms of collective, anyonic degrees of freedom whose entanglement leads to the emergence of the bulk spacetime.Thomas G. Mertens, Joan Simón, Gabriel Wongwork_cycwrprrijgplluxitariaiu7eWed, 23 Nov 2022 00:00:00 GMTQuantum Gravity in 30 Questions
https://scholar.archive.org/work/mllid5xuefabziu5uy4uz7avs4
Quantum gravity is the missing piece in our understanding of the fundamental interactions today. Given recent observational breakthroughs in gravity, providing a quantum theory for what lies beyond general relativity is more urgent than ever. However, the complex history of quantum gravity and the multitude of available approaches can make it difficult to get a grasp of the topic and its main challenges and opportunities. We provide a guided tour of quantum gravity in the form of 30 questions, aimed at a mixed audience of learners and practitioners. The issues covered range from basic motivational and background material to a critical assessment of the status quo and future of the subject. The emphasis is on structural issues and our current understanding of quantum gravity as a quantum field theory of dynamical geometry beyond perturbation theory. We highlight the identification of quantum observables and the development of effective numerical tools as critical to future progress.Renate Loll, Giuseppe Fabiano, Domenico Frattulillo, Fabian Wagnerwork_mllid5xuefabziu5uy4uz7avs4Wed, 23 Nov 2022 00:00:00 GMTSearch for electroweak production of supersymmetric particles in compressed mass spectra with the ATLAS detector at the LHC
https://scholar.archive.org/work/syrsys6kc5cv5ohkityim6kady
Two analyses searching for the production of supersymmetric particles through the electroweak interaction are presented: the chargino search, targeting the pair production of charginos decaying into W bosons and neutralinos, and the displaced track search, looking for charged tracks arising from the decays of higgsinos into pions. These searches target compressed phase spaces, where the mass difference between the next-to-lightest and lightest supersymmetric particle is relatively small. The searches use proton-proton collision data collected at a centre-of-mass energy of 13 TeV with the ATLAS detector at the LHC. In the chargino search, the targeted mass difference between charginos and neutralinos is close to the mass of the W boson. In such phase space, the chargino pair production is kinematically similar to the WW background, making the chargino signal experimentally challenging to be discriminated from the WW background. Machine learning techniques are adopted to separate the supersymmetric signal from the backgrounds. The results exclude chargino masses up to about 140 GeV for mass splittings down to about 100 GeV, superseding the previous results in particularly interesting regions where the chargino pair production could have hidden behind the looking-alike WW background. In the displaced track search, the mass difference between the produced sparticles and the lightest neutralinos goes down to 0.3 GeV. The experimental signature has a low momentum charged track with an origin displaced from the collision point. The results show that the analysis has the sensitivity to exclude different hypotheses for higgsino masses up to 175 GeV if no excess is observed in data. For lower masses, the larger signal cross-section allows to achieve higher significance for different mass splitting scenarios. All these signal hypotheses have not been probed by any existing analysis of LHC data.Eric Ballabenework_syrsys6kc5cv5ohkityim6kadyTue, 22 Nov 2022 00:00:00 GMTPrimordial black holes from D-parity breaking in SO(10) grand unified theory
https://scholar.archive.org/work/xxmiu3sloff6vlr2tm2w7jxic4
The growing evidence of gravitational waves from binary black hole mergers has renewed the interest in study of primordial black holes (PBH). Here we study a mechanism for the formation of PBH from collapse of pseudo-topological domain walls which form out of equilibrium during inflation and then collapse post inflation. We apply the study to domain wall formation due to D-parity embedded in a supersymmetric grand unified theory (GUT) based on SO(10) and compare the abundance of resulting PBH with the existing constraints. Thus the macroscopic relics can then be used to constrain or rule out a GUT, or demand a refinement of the theory of PBH formation in this class of GUTs.Sasmita Mishra, Urjit A. Yajnikwork_xxmiu3sloff6vlr2tm2w7jxic4Tue, 22 Nov 2022 00:00:00 GMTRussian Structuralism and Semiotics in Literary Criticism and its Reception
https://scholar.archive.org/work/mibbm6msg5hudpmuehlrmd2wji
Rainer Grübelwork_mibbm6msg5hudpmuehlrmd2wjiMon, 21 Nov 2022 00:00:00 GMTSpecial Issue: Ludomaterialities
https://scholar.archive.org/work/mzj2yc7oqrhvhjsxatjw7nrqim
Spiele lassen sich nicht auf abstrakte Regelwerke oder narrative Strukturen reduzieren. Sie existieren in und gehen aus unterschiedlichen materiellen Spielkulturen hervor. Sie werden unter Einsatz von Ressourcen und Arbeit designt und produziert, sie benötigen Spielbretter, Würfel oder elektronische Interfaces und sie interagieren mit den Körpern der Spieler:innen ebenso wie mit den sie umgebenden Räumen. Das erste englischsprachige Special Issue der Spiel|Formen untersucht Ludomaterialitäten, d.h. die materiellen Korrelate von Spielen und spielerischen Praktiken, aus einer interdisziplinären Perspektive. Es ist ein Beitrag zum wachsenden Feld der Material Game Studies, das Forschungen zu Themen von Körperlichkeit und Materialität im Kontext von Spielen bündelt. Die Ausgabe ergänzt Übersetzungen bereits publizierter deutschsprachiger Artikel um neue Beiträge und Perspektiven und trägt so dazu bei, den deutschen Game Studies Diskurs an den im angloamerikanischen Raum proklamierten "Material Turn" anschlussfähig zu machen. Mit den gesammelten Beiträgen fokussiert der Band die Bereiche Design und Fanproduktion, Interfaceforschung, Geräte- und Datenverwertung, Medienpädagogik und Kunstgeschichte, adressiert dabei gleichzeitig aber auch explizit analoge Spielgeräte. Die Spannungen zwischen Material, Form und Regeln sowie zwischen Arbeit und Spiel sind zentrale Themen, die sich durch das Heft ziehen.(:Unkn) Unknown, Mediarep.Org, GamesCoopwork_mzj2yc7oqrhvhjsxatjw7nrqimMon, 21 Nov 2022 00:00:00 GMTCoupling Quantum Matter and Gravity
https://scholar.archive.org/work/7f5x67aemfg6rgntclhofwkw3m
In this contribution we deal with several issues one encounters when trying to couple quantum matter to classical gravitational fields. We start with a general background discussion and then move on to two more technical sections. In the first technical part we consider the question how the Hamiltonian of a composite two-particle system in an external gravitational field can be computed in a systematic post-Newtonian setting without backreaction. This enables us to reliably estimate the consistency and completeness of less systematic and more intuitive approaches that attempt to solve this problem by adding 'relativistic effects' by hand. In the second technical part we consider the question of how quantum matter may act as source for classical gravitational fields via the semiclassical Einstein equations. Statements to the effect that this approach is fundamentally inconsistent are critically reviewed.Domenico Giulini, André Großardt, Philip K. Schwartzwork_7f5x67aemfg6rgntclhofwkw3mMon, 21 Nov 2022 00:00:00 GMTHomeless service users' experiences of empowerment and actualisation in homeless service settings
https://scholar.archive.org/work/a64ds4agwndxrdeiz6qkkdg2vi
The stress and severe situational constraints associated with experiences of homelessness can lead to feelings of disempowerment. Although homeless services have the potential to restore service users' empowerment, mainstream homeless service structures continue to reinforce a sense of powerlessness. In this research I combined a theory of empowering services with the capabilities approach to identify features of homeless services that influence service users' empowerment. The project is composed of three studies: a systematic review and narrative synthesis of empowering homeless interventions, a focus group discussion of Irish homeless service users' empowering and actualising experiences, and a semi-structured interview inquiry of service users' central capabilities in eight European countries. Qualitative accounts were collected and compared via thematic analyses. Findings from the systematic review indicated that empowering supported housing, case management, and skills and knowledge acquisition interventions effectively contribute to outcomes aligned with emotional and behavioural empowerment, and their relational antecedents. Themes garnered from the qualitative studies indicate that the physical and affective dimensions of home, service support orientation, and relational well-being and positive community interaction shape service users' empowering and disempowering experiences. Findings from this research identify the organisational and systemic barriers that must be removed to support individual exits from homelessness and highlights service aspects that can be bolstered to improve the empowering experiences of service users. This programme of research produced findings that support policy and service delivery approaches that prioritise housing-led person-centred supports over authoritarian style service environments that undermine the development of service users' empowerment.Branagh R. O'Shaughnessywork_a64ds4agwndxrdeiz6qkkdg2viMon, 21 Nov 2022 00:00:00 GMTTopological Holography: Towards a Unification of Landau and Beyond-Landau Physics
https://scholar.archive.org/work/p6ncec5ppjao5eqex2pjxnxkzy
We outline a holographic framework that attempts to unify Landau and beyond-Landau paradigms of quantum phases and phase transitions. Leveraging a modern understanding of symmetries as topological defects/operators, the framework uses a topological order to organize the space of quantum systems with a global symmetry in one lower dimension. The global symmetry naturally serves as an input for the topological order. In particular, we holographically construct a String Operator Algebra (SOA) which is the building block of symmetric quantum systems with a given symmetry G in one lower dimension. This exposes a vast web of dualities which act on the space of G-symmetric quantum systems. The SOA facilitates the classification of gapped phases as well as their corresponding order parameters and fundamental excitations, while dualities help to navigate and predict various corners of phase diagrams and analytically compute universality classes of phase transitions. A novelty of the approach is that it treats conventional Landau and unconventional topological phase transitions on an equal footing, thereby providing a holographic unification of these seemingly-disparate domains of understanding. We uncover a new feature of gapped phases and their multi-critical points, which we dub fusion structure, that encodes information about which phases and transitions can be dual to each other. Furthermore, we discover that self-dual systems typically posses emergent non-invertible, i.e., beyond group-like symmetries. We apply these ideas to 1+1d quantum spin chains with finite Abelian group symmetry, using topologically-ordered systems in 2+1d. We predict the phase diagrams of various concrete spin models, and analytically compute the full conformal spectra of non-trivial quantum phase transitions, which we then verify numerically.Heidar Moradi, Seyed Faroogh Moosavian, Apoorv Tiwariwork_p6ncec5ppjao5eqex2pjxnxkzyMon, 21 Nov 2022 00:00:00 GMTNo. 1 (2022): World Scientific Reports
https://scholar.archive.org/work/ofdbwz4ivrfnfap535di2ubwcq
Proceedings of the 1st International Scientific Conference «World Scientific Reports» (November 17-18, 2022). Paris, France, 2022Varius Authorswork_ofdbwz4ivrfnfap535di2ubwcqSun, 20 Nov 2022 00:00:00 GMTUnifying State and Policy-Level Explanations for Reinforcement Learning
https://scholar.archive.org/work/rb2mszf7qvaqvmfezuyys5ak4i
Reinforcement learning (RL) is able to solve domains without needing to learn a model of the domain dynamics. When coupled with a neural network as a function approximator, RL systems can solve complex problems. However, verifying and predicting RL agent behavior is made difficult by these same properties; a learned policy conveys "what" to do, but not "why." This thesis focuses on producing explanations for deep RL, summaries of behavior and their causes that can be used for downstream analysis. Specifically, we focus on the setting where the final policy is obtained from a limited, known set of interactions with the environment. We categorize existing explanation methods along two axes: Whether a method explains single-action behavior or policy-level behavior Whether a method provides explanations in terms of state features or past experiences Under this classification, there are four types of explanation methods, and they enable answering different questions about an agent. We introduce methods for creating explanations of these types. Furthermore, we introduce a unified explanation structure that is a combination of all four types. This structure enables obtaining further information about what an agent has learned and why it behaves as it does. First, we introduce CUSTARD, our method for explaining single-action behavior in terms of state features. CUSTARD's explanation is a decision tree representation of the policy. Unlike existing methods for producing such a decision tree, CUSTARD directly learns the tree without approximating a policy after training and is compatible with existing RL techniques. We then introduce APG-Gen, our approach for creating a policy-level behavior explanation in terms of state features. APG-Gen produces a Markov chain over abstract states that enables predicting future actions and aspects of future states. APG-Gen only queries an agent's Q-values, making no assumptions about an agent's decision-making process. We integrate these two methods to produce a Unified Explanation Tree ( [...]Nicholay Topinwork_rb2mszf7qvaqvmfezuyys5ak4iFri, 18 Nov 2022 00:00:00 GMTLectures on modular forms and strings
https://scholar.archive.org/work/rra4lqiauzezpnt4kq5ag3axnu
The goal of these lectures is to present an informal but precise introduction to a body of concepts and methods of interest in number theory and string theory revolving around modular forms and their generalizations. Modular invariance lies at the heart of conformal field theory, string perturbation theory, Montonen-Olive duality, Seiberg-Witten theory, and S-duality in Type IIB superstring theory. Automorphic forms with respect to higher arithmetic groups as well as mock modular forms enter in toroidal string compactifications and the counting of black hole microstates. After introducing the basic mathematical concepts including elliptic functions, modular forms, Maass forms, modular forms for congruence subgroups, vector-valued modular forms, and modular graph forms, we describe a small subset of the countless applications to problems in Mathematics and Physics, including those mentioned above.Eric D'Hoker, Justin Kaidiwork_rra4lqiauzezpnt4kq5ag3axnuFri, 18 Nov 2022 00:00:00 GMTMatter-antimatter asymmetry restrains the dimensionality of neural representations: quantum decryption of large-scale neural coding
https://scholar.archive.org/work/owy6xa2v3bcx3muopa5kiug5hm
Projections from the study of the human universe onto the study of the self-organizing brain are herein leveraged to address certain concerns raised in latest neuroscience research, namely (i) the extent to which neural codes are multidimensional; (ii) the functional role of neural dark matter; (iii) the challenge to traditional model frameworks posed by the needs for accurate interpretation of large-scale neural recordings linking brain and behavior. On the grounds of (hyper-)self-duality under (hyper-)mirror supersymmetry, inter-relativistic principles are introduced, whose consolidation, as spin-geometrical pillars of a network- and game-theoretical construction, is conducive to (i) the high-precision reproduction and reinterpretation of core experimental observations on neural coding in the self-organizing brain, with the instantaneous geometric dimensionality of neural representations of a spontaneous behavioral state being proven to be at most 16, unidirectionally; (ii) a possible role for spinor (co-)representations, as the latent building blocks of self-organizing cortical circuits subserving (co-)behavioral states; (iii) an early crystallization of pertinent multidimensional synaptic (co-)architectures, whereby Lorentz (co-)partitions are in principle verifiable; and, ultimately, (iv) potentially inverse insights into matter-antimatter asymmetry. New avenues for the decryption of large-scale neural coding in health and disease are being discussed.Sofia Karamintziouwork_owy6xa2v3bcx3muopa5kiug5hmFri, 18 Nov 2022 00:00:00 GMTInvestigating the role of social media in the formation of intellectual deviation in the Kingdom of Saudi Arabia
https://scholar.archive.org/work/2qyo26gc5nfnrdpajemfkaujem
Over the last few years, social media have witnessed a huge demand and use by all segments of modern society, especially the younger generation. Such platforms have become tools for exchanging ideas and opinions and forums for discussion. Generations which, in contrast to their predecessors, have been initiated into a society replete with social media. As such, these generations have had social media shape their lives and mould their way of understanding the world around them, and have also been earmarked for the religious terrorism and extremism which has resulted in shocking events such as 9/11, the Arab Spring, and the dawn of violent religious terrorists, such as ISIS (Jenkins, 2011). This research will aim to combine these two facets of modern day life and assess how they influence and interact with one another in the minds of Saudi Arabia's youth. This is done by identifying the role and influence of social media on intellectual deviation: terrorism in Saudi Arabia and understanding its causes and explaining the part that social learning theory has in facilitating the recruitment of individuals through popular social media in the Kingdom of Saudi Arabia (KSA). Social learning theory is effective in showing how relatives and friends are influenced by social media and allows us to determine whether these social platforms contribute to the propagation of terrorism, especially amongst close-groups. A mixed methods approach is used to collect qualitative and quantitative data. The researcher conducted interviews with two official investigators in the Saudi Counter Terrorism Unit (CTU) and interviews with one of the families affected by the terrorist groups. The study also received 250 separate questionnaire replies from the general public by electronic survey to elicit their opinions and attitudes towards the phenomenon of terrorism. In addition, 50 questionnaires were received from investigators in the Saudi Public Prosecution Service and also from experts in criminology at Naif University, KSA. The results of [...]Ibrahim F Al Saabwork_2qyo26gc5nfnrdpajemfkaujemThu, 17 Nov 2022 00:00:00 GMTA Coherent Approach towards Quantum Gravity
https://scholar.archive.org/work/fgscba4m5ffuvbh6aturlwkkjq
This paper typically focuses on the rescaling or equivocally a phase transition from the asymptotic approach of renormalizing the quantum gravity to a more granular approach of the loop quantum gravity (LQG) and then merging it with the Regge calculus for deriving the spin-(2) graviton as the basis of the unified theory. To construct a successful Ultraviolet (UV) completed theory via the fixed-point renormalization group flow equations (FRGE) results in an asymptotic safety approach of the quantum gravity (QG). From the loop-(2) onwards, the higher derivative divergence terms like the higher derivative curvatures, and quadratic divergences with higher derivative scalars make the momentum go to infinity which assaults a problem in renormalizing the QG. If the Einstein-Hilbert (E-H) action, which is the principle of least action is being computed, arising an equation of motion, and a localized path integral (or partition functions) is defined over a curved space, then that action is shown to be associated with the higher order dimension in a more compactified way, resulting in an infinite winding numbers being accompanied through the exponentiality coefficients of the partition integrals in the loop expansions of the second order term onwards, and based on that localization principle, the entire path integral got collapsed to isolated points or granules that if corresponds the aforesaid actions, results in negating the divergences' with an implied bijections' and reverse bijections' of a diffeomorphism of a continuous differentiable functional domains. If those domains are being attributed to the spatial constraints, Hamiltonian constraints, and Master constraints then, through Ashtekar variables, it can be modestly shown that the behavior of quantum origin of asymptotic safety behavior is similar to the LQG granules of spin foam spacetime. Then, we will proceed with the triangulation of the "zoomed in" entangled-points that results in the inclusion of Regge poles via the quantum number (+2,-2,0) as the produced variables of the spin-(2) graviton and spin-(0) dilaton.Deep Bhattacharjeework_fgscba4m5ffuvbh6aturlwkkjqWed, 16 Nov 2022 00:00:00 GMTEssays on the Political Economy of Democratization and Democratic Backsliding
https://scholar.archive.org/work/ls5lve4gtngrfpmv3r5m3id47i
This four-paper dissertation addresses three fundamental questions in the political economy of democracy. First, does economic development cause democratization? Second, to what extent are citizens willing to defend democracy after it has been established, but is then threatened from within by an anti-democratic state executive? Third, what influence has the recent episode of democratic backsliding within the United States had on America's soft power abroad? In the first paper, I provide a new theory of the relationship between economic development and democracy. I argue that a large share of employment in manufacturing (i.e., industrialization) makes mass mobilization both more likely to occur and more costly to suppress. This increases the power of the masses vis-à-vis autocratic elites, making democracy more likely. Using novel manufacturing employment data for 145 countries over 170 years (1845–2015), I find that industrialization is strongly correlated with democracy, even after accounting for country and time fixed effects, time trends, theoretically grounded controls, and other economic determinants of democracy (e.g., income and inequality). Unlike with other economic determinants, the effect occurs on both democratic transitions and consolidations, and is equally large after 1945. Importantly, the data suggests that many potential outliers (e.g., China, the USSR, and Latin America during import substitution industrialization) have in fact never reached the level of industrialization that existed in the West, South Korea, and Taiwan before democratization. In the second paper, I exploit a unique quasi-experiment in 19th- and early 20th-century Norway to test whether the correlation between manufacturing employment and democracy is causal. Using novel roll-call data from the Norwegian national parliament, I study whether MPs that represented more rapidly industrializing districts were more likely to vote for suffrage extensions over the 1891 to 1906 period. For causal identification, I exploit the fact tha [...]Sam Van Noort, Apollo-University Of Cambridge Repository, Ha-Joon Chang, Toke Aidtwork_ls5lve4gtngrfpmv3r5m3id47iWed, 16 Nov 2022 00:00:00 GMTPhysics-Informed Machine Learning: A Survey on Problems, Methods and Applications
https://scholar.archive.org/work/a6ifx4g5tnd37bipbzvdixutke
Recent advances of data-driven machine learning have revolutionized fields like computer vision, reinforcement learning, and many scientific and engineering domains. In many real-world and scientific problems, systems that generate data are governed by physical laws. Recent work shows that it provides potential benefits for machine learning models by incorporating the physical prior and collected data, which makes the intersection of machine learning and physics become a prevailing paradigm. In this survey, we present this learning paradigm called Physics-Informed Machine Learning (PIML) which is to build a model that leverages empirical data and available physical prior knowledge to improve performance on a set of tasks that involve a physical mechanism. We systematically review the recent development of physics-informed machine learning from three perspectives of machine learning tasks, representation of physical prior, and methods for incorporating physical prior. We also propose several important open research problems based on the current trends in the field. We argue that encoding different forms of physical prior into model architectures, optimizers, inference algorithms, and significant domain-specific applications like inverse engineering design and robotic control is far from fully being explored in the field of physics-informed machine learning. We believe that this study will encourage researchers in the machine learning community to actively participate in the interdisciplinary research of physics-informed machine learning.Zhongkai Hao, Songming Liu, Yichi Zhang, Chengyang Ying, Yao Feng, Hang Su, Jun Zhuwork_a6ifx4g5tnd37bipbzvdixutkeTue, 15 Nov 2022 00:00:00 GMTDesign and training of deep reinforcement learning agents
https://scholar.archive.org/work/v4bnexxtdbgkvdgyu4jub7gulm
Deep reinforcement learning is a field of research at the intersection of reinforcement learning and deep learning. On one side, the problem that researchers address is the one of reinforcement learning: to act efficiently. A large number of algorithms were developed decades ago in this field to update value functions and policies, explore, and plan. On the other side, deep learning methods provide powerful function approximators to address the problem of representing functions such as policies, value functions, and models. The combination of ideas from these two fields offers exciting new perspectives. However, building successful deep reinforcement learning experiments is particularly difficult due to the large number of elements that must be combined and adjusted appropriately. This thesis proposes a broad overview of the organization of these elements around three main axes: agent design, environment design, and infrastructure design. Arguably, the success of deep reinforcement learning research is due to the tremendous amount of effort that went into each of them, both from a scientific and engineering perspective, and their diffusion via open source repositories. For each of these three axes, a dedicated part of the thesis describes a number of related works that were carried out during the doctoral research. The first part, devoted to the design of agents, presents two works. The first one addresses the problem of applying discrete action methods to large multidimensional action spaces. A general method called action branching is proposed, and its effectiveness is demonstrated with a novel agent, named BDQ, applied to discretized continuous action spaces. The second work deals with the problem of maximizing the utility of a single transition when learning to achieve a large number of goals. In particular, it focuses on learning to reach spatial locations in games and proposes a new method called Q-map to do so efficiently. An exploration mechanism based on this method is then used to demonstrate the effect [...]Fabio Pardo, Petar Kormushev, Andrew Davison, Dyson Technology Limited (Firm)work_v4bnexxtdbgkvdgyu4jub7gulmTue, 15 Nov 2022 00:00:00 GMT