Angelo Dalli (born 14 April 1978) is a computer scientist specialising in artificial intelligence, a serial entrepreneur, and business angel investor. == Early life and education == Dalli was born in Malta and grew up in the town of Birżebbuġa. Dalli was educated at the Archbishop's Seminary, Malta and represented Malta in the Young European Environmental Research contest held in Cologne in 1994. Dalli represented Malta in the International Olympiad in Informatics held in Eindhoven in 1995, where he won a bronze medal. Dalli started selling computer software as a teenager, and worked for the International Data Group as a freelance contributor for PC World. == Academic work == After graduating from the University of Malta, Dalli spent time lecturing on artificial intelligence and natural language processing before reading for his PhD at the University of Sheffield under the supervision of Yorick Wilks. Dalli has published over 23 peer reviewed papers in the artificial intelligence and natural language processing fields, including one of the earliest methods on timestamp extraction from documents that is now commonly used in most email applications. Angelo has also contributed to the encoding of European languages in Unicode, in particular for the Common Locale Data Repository. In the field of Bioinformatics Dalli has found a particularly useful integer sequence (sequence A062208 in the OEIS) which efficiently computes all alignments of strings of length 3 together with other generalisations (sequence A062204 in the OEIS), (sequence A062205 in the OEIS) for applications in natural language and sequence alignment. Dalli has an Erdős number of 3. Dalli has led the Maltese national informatics team in the International Olympiad in Informatics at IOI 2002 in Seoul, South Korea and IOI 2004 in Athens, Greece. == Artificial intelligence == === Trustworthy AI and Hybrid Intelligence === Angelo has been a vocal proponent of trustworthy AI that impacts society positively and believes that AI should be properly regulated. Angelo has co-founded UMNAI in 2019, with the aim of creating a new form of trustworthy AI that can explain the decisions and steps that the AI has taken to output an answer, based on a neurosymbolic AI architecture that combines neural and symbolic AI in an auditable and certain manner. === AI and society === Angelo led the Government of Malta taskforce that produced Malta's new AI regulation and national AI strategy, and is an active member of the IEEE, AAAI, ACM and the ACL. === AI in transport === Angelo had led the introduction of different machine learning techniques in intelligent transport systems (ITS), including parking, controlled vehicle access zones and dynamic traffic interchange control. His intelligent transport company, Traffiko, operated in Europe, Australia and the Middle East, and was eventually sold to Q-Free in Norway in 2015. === AI in gaming === Angelo is a well known speaker in the online gambling industry. Angelo setup one of the first companies that applied artificial intelligence in the online gambling industry, called Bit8 (now part of Intralot), with the most notable work being on algorithms that estimate and maximise player lifetime value and personalised bonusing systems. These techniques have since been widely adopted by the online gambling industry Intralot subsequently bought Bit8 in 2017. === AI and creativity === Angelo has been collaborating various artists and creatives to teach AI about creativity. The results of this collaboration is the UMA AI entity, short for Universal Machine Artist. Angelo has also co-founded the Creative Science and Arts Institute to act as a foundation for future research into AI, science, technology and creativity. UMA is creating original artwork using a modified Generative adversarial network has a third component, the human artist, to produce different learning results than standard generative AI models. The underlying discriminator in UMA started from an anti-fraud detection system and has now gradually evolved to add stable diffusion and procedural generation methods. The first two artworks generated by UMA were auctioned in October and November 2018 respectively, with all proceeds donated to charity and good causes. Ongoing work in improving UMA and furthering collaboration with other artists is ongoing. Notable exhibitions include Tomorrow's Blossoms with Selina Scerri at Esplora Museum in 2024, which explored the theme of AI and emotions. == Angel investor == Angelo is an angel investor active in the high-tech startup scene, and is a member of EBAN, and World Business Angel Forum senator. Angelo has been encouraging Maltese startups via various public events including the Zest and Budding Rockstars conferences and co-founded BAM, the Malta Business Angel network, in 2019. == Awards and honours == === Entrepreneurial and scientific === Bronze Medal, International Olympiad in Informatics (1995) Malta Top Entrepreneur Award (2019) Malta Top Entrepreneur Award (2014) WIPO IP Enterprise Award for the UMNAI Neuro-symbolic AI architecture (2022) === Corporate awards === Intralot Bit8 EGR Rising Star Award (2014) Intralot Bit8 Malta Communication Authority eBusiness Award for the Best B2B application (2015) Intralot Bit8 Malta iGaming Award for Excellence (2017)
Harmony (software)
Harmony is a Java-based software for creating high-definition music videos with 2D and 3D animations. The application was developed by Digital Chaotics, a company based in San Jose, California and established in 2010 by Ken and Leanna Scott. == History == During a March 1, 2011 interview published by The LIST magazine, Ken explained how he initially got into music and digital entertainment. According to Scott: “I came at it from both the art and the technology side. … I built one of the first digital audio synthesizers as an undergrad project back in 1979. It was a short jump from there to creating visuals with computers, too.” Taking inspiration from Fantasia – which Scott calls, “The greatest music video of all time” – he began writing software code for Harmony in late 2009, finishing the project in mid-2010. However, Scott has also said that the idea for Harmony began much earlier: I read a book in 1978 called Digital Harmony, by John H Whitney, Sr. (Interestingly, he was the father of the president of Digital Productions.) He said that there was a kind of visual art based on motion, and proposed theories about the underlying mathematical structure of visual harmony. So there's the book, combined with my desire to create art with computers-add a taste or two of things commonly used by college students during the 70's - and lots of Pink Floyd. Add it all up, and the seeds for Harmony were planted. My friends in school and at Floating Point Systems listened to me ranting about "making music videos with computers" incessantly. I'm sure it was both maddening and fascinating to see. == Features == Harmony runs on Windows 7 and Windows Vista. Currently, Digital Chaotics does not offer a macOS or Linux platform for the software. However, Harmony can be run on these platforms by running it on Windows in a virtual machine. == Harmony 2 == On November 1, 2011, Digital Chaotics released the 2.0 version of the Harmony software. Unlike the original version, the second release featured three product levels: Harmony 2 Express, Harmony 2 Pro, and Harmony 2 Extreme. The "Express" version was positioned as an entry-level, free release to allow users a chance to "test-drive" the software. The "Pro" version currently retails at $197, while the "Extreme" is priced at $397. These two versions, aimed more towards VJ and Fulldome theater usage, featured additional software capability and features such as higher resolution, more video formatting options, and more camera angles.
Informetrics
Informetrics is the study of quantitative aspects of information, it is an extension and evolution of traditional bibliometrics and scientometrics. Informetrics uses bibliometrics and scientometrics methods to study mainly the problems of literature information management and evaluation of science and technology. Informetrics is an independent discipline that uses quantitative methods from mathematics and statistics to study the process, phenomena, and law of informetrics. Informetrics has gained more attention as it is a common scientific method for academic evaluation, research hotspots in discipline, and trend analysis. Informetrics includes the production, dissemination, and use of all forms of information, regardless of its form or origin. Informetrics encompasses the following fields: Scientometrics, which studies quantitative aspects of science Webometrics, which studies quantitative aspects of the World Wide Web Bibliometrics, which studies quantitative aspects of recorded information Cybermetrics, which is similar to webometrics, but broadens its definition to include electronic resources == Origin and Development == The term informetrics (French: informétrie) was coined by German scholar Otto Nacke in 1979, and came from the German word 'informetrie’. The corresponding English terminology soon appeared in the subsequent literature. In September 1980, Professor Otto Nacke introduced the term 'informetrics' at the first seminar on Informetrics in Frankfurt, Germany. Later, Committee on Informetrics has established through The International Federation for Information and Documentation (FID). In 1987, informetrics started to be officially recognized by the international information community and several foreign information scientists. In 1988, at First International Conference on Bibliometrics and Theoretical Aspects of Information Retrieval Archived 2022-05-23 at the Wayback Machine, Brooks suggested bibliometrics and scientometrics can be included in the field of informetrics. In 1990, Leo Egghe and Ronald Rousseau proposed the formation of the discipline of informetrics: statistical bibliography (1923) to bibliometrics and scientometrics (1969) and then to informetrics (1979). In 1993, the International Society for Scientometrics and Informetrics (ISSI) Archived 2023-11-05 at the Wayback Machine was founded at the International Conference on Bibliometrics, Informetrics and Scientometrics in Berlin, and the first one was held in Belgium and organized by Leo Egghe and Ronald Rousseau. The society was formally incorporated in 1994 in the Netherlands and plays a significant role in the development of informetrics. The ISSI aims to promote the "exchange and communication of professional information in the fields of scientometrics and informetrics, including improve standards, theory and practice, as well as promote research, education and training". In addition, to "engage in relevant public conversation and policy discussions". In the western world, 20th century's Informetrics is mostly based on Lotka's law, named after Alfred J. Lotka, Zipf's law, named after George Kingsley Zipf, Bradford's law named after Samuel C. Bradford and on the work of Derek J. de Solla Price, Gerard Salton, Leo Egghe, Ronald Rousseau, Tibor Braun, Olle Persson, Peter Ingwersen, Manfred Bonitz, and Eugene Garfield. == Difference Between Informetrics, Bibliometrics and Scientometrics == Since the 1960s, three similar terms have emerged in the fields of library science, philology and science of science, they are bibliometrics, scientometrics and informetrics, representing three very similar quantitative sub-disciplines. The three metrics terms can be confusing and often misused. Informetrics and bibliometrics interpenetrate each other but have different aspects in research object, research scope, and measuring unit. Informetrics and scientometrics are very different in their research purpose and research object, as well as the research scope and application. Bibliometrics is categorised under the field of library science, it uses mathematical and statistical methods to describe, evaluate, and predict the current status and trends of science and technology. Also to study the "distribution structure, quantitative relationship, change law and quantitative management of literature information, quantitative relationships, patterns and quantitative management of literature and information". The term was first used by Alan Pritchard in 1969 in his paper Statistical Bibliography or Bibliometrics?. Scientometrics is a branch of science that quantitatively evaluates and predicts the process and management of scientific activities in order to reveal their development patterns and trends. The definition of scientometrics was described by Derek De Solla Price in his book Science to Science as the “quantitative study of science, communication in science, and science policy”. === Links between the three metrics terms === The most prominent connection between the three metrics terms is in their research objects. Since all three disciplines use literature information as their research object, therefore, they have some similarities and overlaps in their research methods and fields. Moreover, they all use mathematical methods as the basic research methods and they all apply the three basic laws, Bradford's law, Lotka's law and Zipf's law. === Distinctions between the three metrics terms === The distinction between the three metrics terms can tell from their research object and research purpose. The research of bibliometrics focuses on the analysis of "scientific output in the form of articles, publications, citations, and others". Scientometrics is to measure the basic characteristics and laws of scientific activities. Where informetrics is to investigate information sources and information distribution process. == Concept and System Structure == === Purpose of Informetrics Research === The main purpose of informetrics is to use its theocratical research to solve the methodological issues in the research process, and to discover and reveal the basic laws of information distribution through the study of information process and phenomenon. In this way, makes information management more scientific and provides a quantitative basis for information services and information management decisions. For informetrics, it is necessary to bring quantitative analysis methods to further reveal the structure of information units and the "quantitative change law of literature information”. Further to this, to improve the scientific accuracy of information science from a theoretical point of view. At the same time, to better solve the basic contradictions in the information service, overcome the information crisis, and make the information management work more effective to serve science and technology, economic and social development. Quantitative analysis of bibliographic data was pioneered by Robert K. Merton in an article called Science, Technology, and Society in Seventeenth Century England and originally published by Merton in 1938. === The Significance of Informetrics Research === The significance of informetrics research is to summarize various empirical laws from the theoretical point of view, at the same time test and modify the various empirical laws in the new information unit conditions, and explore its new applicability, therefore, the scientific nature of information science can be improved, but also to provide theoretical guidance for practical work. === The Objects of Informetrics Research === The object of informetrics is broader than the field of bibliometrics and scientometrics, including "messages, data, events, objects, text, and documents”. Informetrics is often used to inform policies and decisions across a broad range of fields, such as economy, politics, technology and social spheres that "influence the flow and use patterns of information". Tague-Sutcliffe describes the following uses of informetrics: Citation analysis; Characteristics of authors; Use of recorded information; Obsolescence of the literature; Concomitant growth of new concepts; Characteristics of publication sources; Definition and measurement o information; Growth of subject literature, databases, libraries; Types and characteristics of retrieval performance measures; Statistical aspects of language, word, and phrase frequencies. == Basic Laws == In the field of informetrics research, there are many outstanding contributors in the discipline with a solid knowledge of quantitative research methods. In the early 20th century, several scientists contributed empirical applications that have become the three basic laws of informetrics, Bradford's law, Lotka's law, and Zipf's law, which promote the development of informetrics. === Bradford's Law === The British documentalist and librarian Samuel C. Bradford first discovered the law of concentration and scattering of literature, and in 1934, it has be
Exploratory search
Exploratory search is a specialization of information exploration which represents the activities carried out by searchers who are: unfamiliar with the domain of their goal (i.e. need to learn about the topic in order to understand how to achieve their goal) or unsure about the ways to achieve their goals (either the technology or the process) or unsure about their goals in the first place. Exploratory search is distinguished from known-item search, for which the searcher has a particular target in mind. Consequently, exploratory search covers a broader class of activities than typical information retrieval, such as investigating, evaluating, comparing, and synthesizing, where new information is sought in a defined conceptual area; exploratory data analysis is another example of an information exploration activity. Typically, therefore, such users generally combine querying and browsing strategies to foster learning and investigation. == History == Exploratory search is a topic that has grown from the fields of information retrieval and information seeking but has become more concerned with alternatives to the kind of search that has received the majority of focus (returning the most relevant documents to a Google-like keyword search). The research is motivated by questions like "What if the user doesn't know which keywords to use?" or "What if the user isn't looking for a single answer?" Consequently, research has begun to focus on defining the broader set of information behaviors in order to learn about the situations when a user is, or feels, limited by only having the ability to perform a keyword search. In the last few years, a series of workshops has been held at various related and key events. In 2005, the Exploratory Search Interfaces workshop focused on beginning to define some of the key challenges in the field. Since then a series of other workshops has been held at related conferences: Evaluating Exploratory Search at SIGIR06 and Exploratory Search and HCI at CHI07 (in order to meet with the experts in human–computer interaction). In March 2008, an Information Processing and Management special issue focused particularly on the challenges of evaluating exploratory search, given the reduced assumptions that can be made about scenarios of use. In June 2008, the National Science Foundation sponsored an invitational workshop to identify a research agenda for exploratory search and similar fields for the coming years. == Research challenges == === Important scenarios === With the majority of research in the information retrieval community focusing on typical keyword search scenarios, one challenge for exploratory search is to further understand the scenarios of use for when keyword search is not sufficient. An example scenario, often used to motivate the research by mSpace, states: if a user does not know much about classical music, how should they even begin to find a piece that they might like. Similarly, for patients or their carers, if they don't know the right keywords for their health problems, how can they effectively find useful health information for themselves? === Designing new interfaces === With one of the motivations being to support users when keyword search is not enough, some research has focused on identifying alternative user interfaces and interaction models that support the user in different ways. An example is faceted search which presents diverse category-style options to the users, so that they can choose from a list instead of guess a possible keyword query. Many of the interactive forms of search, including faceted browsers, are being considered for their support of exploratory search conditions. Computational cognitive models of exploratory search have been developed to capture the cognitive complexities involved in exploratory search. Model-based dynamic presentation of information cues are proposed to facilitate exploratory search performance. === Evaluating interfaces === As the tasks and goals involved with exploratory search are largely undefined or unpredictable, it is very hard to evaluate systems with the measures often used in information retrieval. Accuracy was typically used to show that a user had found a correct answer, but when the user is trying to summarize a domain of information, the correct answer is near impossible to identify, if not entirely subjective (for example: possible hotels to stay in Paris). In exploration, it is also arguable that spending more time (where time efficiency is typically desirable) researching a topic shows that a system provides increased support for investigation. Finally, and perhaps most importantly, giving study participants a well specified task could immediately prevent them from exhibiting exploratory behavior. === Models of exploratory search behavior === There have been recent attempts to develop a process model of exploratory search behavior, especially in social information system (e.g., see models of collaborative tagging. The process model assumes that user-generated information cues, such as social tags, can act as navigational cues that facilitate exploration of information that others have found and shared with other users on a social information system (such as social bookmarking system). These models provided extension to existing process model of information search that characterizes information-seeking behavior in traditional fact-retrievals using search engines. Recent development in exploratory search is often concentrated in predicting users' search intents in interaction with the user. Such predictive user modeling, also referred as intent modeling, can help users to get accustomed to a body of domain knowledge and help users to make sense of the potential directions to be explored around their initial, often vague, expression of information needs. == Major figures == Key figures, including experts from both information seeking and human–computer interaction, are: Marcia Bates Nicholas Belkin Gary Marchionini m.c. schraefel Ryen W. White
National Data Repository
A National Data Repository (NDR) is a data bank that seeks to preserve and promote a country's natural resources data, particularly data related to the petroleum exploration and production (E&P) sector. A National Data Repository is normally established by an entity that governs, controls and supports the exchange, capture, transference and distribution of E&P information, with the final target to provide the State with the tools and information to assure the growth, govern-ability, control, independence and sovereignty of the industry. The two fundamental reasons for a country to establish an NDR are to preserve data generated inside the country by the industry, and to promote investments in the country by utilizing data to reduce the exploration, production, and transportation business risks. Countries take different approaches towards preserving and promoting their natural resources data. The approach varies according to a country's natural resources policies, level of openness, and its attitude towards foreign investment. == Data types == NDRs store a vast array of data related to a country's natural resources. This includes wells, well log data, well reports, core samples, seismic surveys, post-stack seismic, field data/tapes, seismic (acquisition/processing) reports, production data, geological maps and reports, license data and geological models. == Funding models == Some NDRs are financed entirely by a country's government. Others are industry-funded. Still some are hybrid systems, funded in part by industry and government. NDRs typically charge fees for data requests and for data loading. The cost differs significantly between countries. In some cases an annual membership is charged to oil companies to store and access the data in the NDR. == Standards body == Energistics is the global energy standards resource center for the upstream oil and gas industry. Energistics National Data Repository Work Group: The standards body is Energistics. === Energistics-standards-directory === Global regulators of upstream oil and natural gas information, including seismic, drilling, production and reservoir data, formed the National Data Repository (NDR) Work Group in 2008 to collaborate on the development of data management standards and to assist emerging nations with hydrocarbon reserves to better collect, maintain and deliver oil and gas data to the public and to the industry. Ten countries, led by the Netherlands, Norway and the United Kingdom, formed NDR to share best practices and to formalize the development and deployment of data management standards for regulatory agencies. The other countries involved in the NDR Work Group's formation are Australia, Canada, India, Kenya, New Zealand, South Africa and the United States. Annual NDR Conference: Approximately every 18 months Energistics organizes a National Data Repository Conference. The purpose is to provide government and regulatory agencies from around the world an opportunity to attend a series of workshops dedicated to developing data exchange standards, improving communications with the oil and gas industry and learning data management techniques for natural resources information. === Society of Exploration Geophysicists and The International Oil and Gas Producers Association === The SEG is the custodian of the SEG standards which are used for the exchange, retention and release of seismic data. They are commonly used by National Data Repositories with the SEGD and SEGY being the field and processed exchange standards respectively. == NDRs around the world == Click here to see a map of the NDRs around the world
List of assembly software and tools
This is a list of assembly software and tools, including software used for assembly language programming, machine code generation, disassembly, debugging, binary analysis, reverse engineering, and instruction-set simulation. == Assemblers and machine-code generators == == Disassemblers and binary-analysis tools == == Debuggers with assembly-level features == == Educational IDEs, simulators and emulators == == Portable and intermediate assembly-like languages == == Assembly language families == Assembly language is not a single programming language, but a family of low-level languages associated with particular instruction set architectures and processor families. Examples include:
Lai–Robbins lower bound
The Lai–Robbins lower bound gives an asymptotic lower bound on the regret that any uniformly good algorithm must incur in the stochastic multi-armed bandit problem. The original result was proved by Tze Leung Lai and Herbert Robbins in 1985 for parametric exponential families. Later work extended the statement to more general classes of distributions. == Multi-armed bandit problem == The multi-armed bandit problem (MAB) is a sequential game in which the player must trade off exploration (to learn) and exploitation (to earn). The player chooses among K {\displaystyle K} actions (arms) with unknown distributions ν = ( ν 1 , … , ν K ) {\displaystyle \nu =(\nu _{1},\dots ,\nu _{K})} . The player is assumed to know a class of distributions D {\displaystyle {\mathcal {D}}} such that for every k {\displaystyle k} one has ν k ∈ D {\displaystyle \nu _{k}\in {\mathcal {D}}} (for example, D {\displaystyle {\mathcal {D}}} may be the family of Gaussian or Bernoulli distributions). At each round t = 1 , … , T {\displaystyle t=1,\dots ,T} the player selects (pulls) an arm a t {\displaystyle a_{t}} and observes a reward X t ∼ ν a t {\displaystyle X_{t}\sim \nu _{a_{t}}} . We denote N a ( t ) := ∑ s = 1 t 1 { a s = a } {\displaystyle N_{a}(t):=\sum _{s=1}^{t}\mathbf {1} _{\{a_{s}=a\}}} the number of times arm a {\displaystyle a} has been pulled in the first t {\displaystyle t} rounds, μ ( ν ) := ( μ 1 , … , μ K ) {\displaystyle \mu (\nu ):=(\mu _{1},\dots ,\mu _{K})} the vector of arm means, where μ k = E X ∼ ν k [ X ] {\displaystyle \mu _{k}=\mathbb {E} _{X\sim \nu _{k}}[X]} , μ ∗ := max a μ a {\displaystyle \mu ^{}:=\max _{a}\mu _{a}} the highest mean Δ a := μ ∗ − μ a ≥ 0 {\displaystyle \Delta _{a}:=\mu ^{}-\mu _{a}\geq 0} the gap of arm a {\displaystyle a} . An arm a {\displaystyle a} with μ a = μ ∗ {\displaystyle \mu _{a}=\mu ^{}} is called an optimal arm; otherwise it is a suboptimal arm. The goal is to minimize the regret at horizon T {\displaystyle T} , defined by R T := ∑ a = 1 K Δ a E [ N a ( T ) ] . {\displaystyle R_{T}:=\sum _{a=1}^{K}\Delta _{a}\,\mathbb {E} [N_{a}(T)].} Intuitively, the regret is the (expected) total loss compared to always playing an optimal arm: regret = ∑ a ( cost of playing a ) × ( times a is played ) . {\displaystyle {\text{regret}}=\sum _{a}\ ({\text{cost of playing }}a)\times ({\text{times }}a{\text{ is played}}).} An MAB algorithm is a (possibly randomized) policy that, at each round t {\displaystyle t} , choose an arm a_t by using the observations received from previous turns. === Intuitive example === Suppose a farmer must choose, each year, one of K {\displaystyle K} seed varieties to plant. Each variety k {\displaystyle k} has an unknown average yield μ k {\displaystyle \mu _{k}} . If the farmer knew the best variety (with mean μ ∗ {\displaystyle \mu ^{}} ) he would plant it every year; in reality he must try varieties to learn which is best. The cumulative regret after T {\displaystyle T} years measures the total expected loss in yield due to imperfect knowledge. Remarks The model above is the stochastic MAB; there also exist adversarial variants. One may consider a fixed-horizon setting (known T {\displaystyle T} ) or an anytime setting (unknown T {\displaystyle T} ). == Lai–Robbins lower bound == The theorem gives the right amount of time we should pull a suboptimal arm k {\displaystyle k} to distinguish whether we are in the instance with ν k {\displaystyle \nu _{k}} or with ν ~ k {\displaystyle {\tilde {\nu }}_{k}} where ν ~ k {\displaystyle {\tilde {\nu }}_{k}} is such that μ ~ k > μ ∗ {\displaystyle {\tilde {\mu }}_{k}>\mu ^{}} . Knowning a lower bound on the number of pull of every suboptimal arm gives a lower bound on the regret as only suboptimal arms contribute to the regret. Before stating the formal theorem we need to define what is a consistent algorithm. === Consistency (uniformly good algorithms) === Let D {\displaystyle {\mathcal {D}}} be a class of probability distributions and consider K {\displaystyle K} arms with reward distributions ν = ( ν 1 , … , ν K ) ∈ D K {\displaystyle \nu =(\nu _{1},\dots ,\nu _{K})\in {\mathcal {D}}^{K}} . An algorithm is said to be consistent (also called uniformly good) on D K {\displaystyle {\mathcal {D}}^{K}} if, for every instance ν ∈ D K {\displaystyle \nu \in {\mathcal {D}}^{K}} , the expected regret R T ( ν ) {\displaystyle R_{T}(\nu )} grows subpolynomially: ∀ α > 0 , R T ( ν ) = o ( T α ) as T → ∞ {\displaystyle \forall \alpha >0,\qquad R_{T}(\nu )=o(T^{\alpha })\quad {\text{as }}T\to \infty } This assumption excludes algorithms that perform well on some instances but incur linear regret on others. === Formal lower bound === For any suboptimal arm a {\displaystyle a} . For a distribution ν a ∈ D {\displaystyle \nu _{a}\in {\mathcal {D}}} and a threshold x {\displaystyle x} , define K inf ( ν a , x , D ) := inf { KL ( ν a , ν ′ ) : ν ′ ∈ D , μ ′ > x } {\displaystyle {\mathcal {K}}_{\inf }(\nu _{a},x,{\mathcal {D}}):=\inf {\Bigl \{}\operatorname {KL} (\nu _{a},\nu '):\nu '\in {\mathcal {D}},\ \mu '>x{\Bigr \}}} where KL ( ⋅ , ⋅ ) {\displaystyle \operatorname {KL} (\cdot ,\cdot )} denotes the Kullback-Leibler divergence. Then, for any algorithm consistent on D K {\displaystyle {\mathcal {D}}^{K}} and for every instance ν ∈ D K {\displaystyle \nu \in {\mathcal {D}}^{K}} , every suboptimal arm a {\displaystyle a} satisfies E ν [ N a ( T ) ] ≥ ln T K inf ( ν a , μ ∗ , D ) + o ( ln T ) {\displaystyle \mathbb {E} _{\nu }[N_{a}(T)]\geq {\frac {\ln T}{{\mathcal {K}}_{\inf }(\nu _{a},\mu ^{},{\mathcal {D}})}}+o(\ln T)} Consequently, the regret satisfies R T ( ν ) ≥ ( ∑ a : μ a < μ ∗ Δ a K inf ( ν a , μ ∗ , D ) ) ln T + o ( ln T ) {\displaystyle R_{T}(\nu )\geq \left(\sum _{a:\,\mu _{a}<\mu ^{}}{\frac {\Delta _{a}}{{\mathcal {K}}_{\inf }(\nu _{a},\mu ^{},{\mathcal {D}})}}\right)\ln T+o(\ln T)} The original 1985 paper established this result for exponential families; later work showed that the bound holds under much weaker assumptions on D {\displaystyle {\mathcal {D}}} . === Intuition === Consistency imposes that, for every ν {\displaystyle \nu } , the number of pulls of an optimal arm must be large. This means that μ ∗ {\displaystyle \mu ^{}} is estimated very accurately. The goal is to determine, for a suboptimal arm k {\displaystyle k} , how many samples are needed to be confident, with the appropriate level of confidence, that μ k < μ ∗ {\displaystyle \mu _{k}<\mu ^{}} . To do so, we use what is called the most confusing instance: an instance close to ν {\displaystyle \nu } such that arm k {\displaystyle k} is optimal. We define it as ν ~ {\displaystyle {\tilde {\nu }}} such that, for all a ≠ k {\displaystyle a\neq k} , ν ~ a = ν a {\displaystyle {\tilde {\nu }}_{a}=\nu _{a}} , and ν ~ k {\displaystyle {\tilde {\nu }}_{k}} is chosen so that μ ~ k > μ ∗ {\displaystyle {\tilde {\mu }}_{k}>\mu ^{}} . The objective is to determine how many samples of arm k {\displaystyle k} are required to distinguish whether we are in the instance with ν k {\displaystyle \nu _{k}} or with ν ~ k {\displaystyle {\tilde {\nu }}_{k}} in terms of KL {\displaystyle \operatorname {KL} } distance. == Algorithms achieving the Lai–Robbins lower bound == Several algorithms are known to achieve the Lai–Robbins asymptotic lower bound under specific assumptions on the reward distribution class D {\displaystyle {\mathcal {D}}} . The following list summarizes a non-exhaustive list of algorithms matching the lower bound. == Extension to other problems == === Structured bandit === A more complexe is structured bandit where we know that the mean of each arm is in a set with some restriction. In this case we can prove a smaller lower bound that use the knowledge of this set. === Best arm identification (BAI) === A similar result has been proved for best arm identification, which is the same game except that, instead of minimizing the regret, the goal is to identify the best arm with probability 1 − δ {\displaystyle 1-\delta } using as few rounds as possible. === Reinforcement Learning (RL) === Similar results have been proved for regret minimization in average-reward reinforcement learning. The order is also ln T {\displaystyle \ln T} , with a constant that depends on the problem.