Festival of International Virtual & Augmented Reality Stories (FIVARS) is a Canadian media festival for story-driven works using extended reality (XR) and immersive media, including virtual reality, augmented reality, WebXR, live VR performance, projection mapping and spatialized audio. Founded in Toronto in 2015, it has been described as Canada's first dedicated virtual and augmented reality stories festival, the first Canadian festival of its kind, and Canada's original festival dedicated to immersive storytelling. FIVARS has described itself as "the original and longest-running festival wholly dedicated to Virtual and Augmented Reality Stories", while third-party XR coverage has called it one of the longest-running events dedicated to immersive content. FIVARS is produced by Constant Change Media Group, Inc., with its partner event VRTO. == History == FIVARS began in 2015, with preview screenings at the Camp Wavelength music festival on Toronto Island and an inaugural festival held in Toronto in September 2015. Contemporary coverage described the first edition as a virtual reality film festival held at UG3 Live in Toronto. The festival continued with a second edition in 2016. L'Express described the 2016 festival as presenting Canadian and international interactive works in virtual and augmented reality narrative forms. FIVARS's 2016 festival was also listed in a York University Future Cinema course page as a public event students could attend. In 2017, the third annual FIVARS festival was held at the House of VR in Toronto. In 2018, the festival was held at the Matador Ballroom, which NOW Magazine reported was reopening for FIVARS from September 14 to 16. The festival's own history states that the 2018 edition included 36 works from 12 countries and that Stephanie Greenall took over as co-producer that year. In 2019, FIVARS moved to the Toronto Media Arts Centre for its fifth anniversary and listed official selections in passive and interactive immersive-experience categories. The festival also held talks and panels at the Toronto Media Arts Centre. During the COVID-19 pandemic, FIVARS moved part of its programming online. In 2020, Voices of VR reported that Malicki-Sanchez and WebXR developer James Baicoianu used JanusXR code to create a platform for presenting 360-degree video through the web. The festival's history states that its 2020 online festival included 39 selections from 16 countries and was produced by Malicki-Sanchez and Greenall. In 2021, FIVARS introduced a dual-event structure with FIVARS in FEB and FIVARS in FALL. The fall 2021 edition used a hybrid format, with an in-person component in West Hollywood from October 15 to 17 and an online WebXR component from October 22 to November 2. In 2022, FIVARS held hybrid programming with pop-up viewing locations in Los Angeles and Toronto. The fall 2022 edition was listed by blogTO as the festival's tenth edition, with an in-person component at Stackt - an outdoor arts park built from shipping containers in Toronto and online programming. The 2023 festival was presented as a hybrid exhibition of 65 immersive stories, with an in-person Toronto component and an online component. The FIVARS Online Festival was later listed among the Innovator of the Year nominees for the 2024 Poly Awards. FIVARS stated that the nominees for that recognition were producer and designer Keram Malicki-Sanchez and developer James Baicoianu. The 2024 edition was listed as FIVARS 2024 (Toronto + Online), with an in-person Toronto event from October 3 to 8 and an online component beginning October 10. The festival also published a 2024 official selections list covering virtual reality, augmented reality, spherical video, spatial web and related immersive formats. In 2025, FIVARS and VRTO were held together at OCAD University. The 2026 edition is scheduled for June 15 to 19, 2026, at OCAD University in Toronto, with OCAD University as presenting sponsor and first-time venue host. FIVARS has featured official selections from more than forty countries across six continents. == Organization == FIVARS was founded in 2015 by Keram Malicki-Sánchez. Joseph Ellsworth was the festival's original technical director and helped operate FIVARS during its early years. Malicki-Sánchez remains executive director and festival director. Jessy Blaze joined Malicki-Sánchez as co-producer in 2016 and served until Stephanie Greenall took over the role in 2018. Greenall served as co-producer and associate producer from 2018 to 2022. Aimee Reynolds took over from Greenall in 2022 and has served as associate producer of FIVARS and VRTO since 2022. == Immersive Media Awards == FIVARS presents People's Choice awards for interactive works and immersive video or passive immersive works. Juried award categories have included the Grand Jury Prize, Impact Award, Technical Achievement, Excellence in Experience Design, Excellence in Visual Design, Excellence in Sound Design, and Outstanding Performance. === 2015 === On Monday, September 21, the festival announced People's Choice awards for two categories at the Cadillac Lounge, a music venue and restaurant in Toronto. People's Choice Best Interactive Experience: Apollo 11 Best Immersive Video: SONAR === 2016 === People's Choice Best Interactive Experience: Pearl (Patrick Osborne) Best Immersive Video: Help (Justin Lin) Juried Grand Jury Award: Real (Connor Hair and Alex Meader) === 2017 === People's Choice Best Interactive: Alteration Best Immersive (Passive): Guardian of the Guge Kingdom Juried Impact Award: Priya's Shakti / Priya's Mirror (Dan Goldman) Grand Jury Prize: Manifest 99 === 2018 === People's Choice Best Interactive: Museum of Symmetry (Paloma Dawkins) Best Immersive (Passive): Going Home (David Beier) Juried Impact Award: The Hidden (Annie Lukowski, BJ Schwartz) Grand Jury Prize: Battlescar (Nico Casavecchia, Martin Allais) === 2019 === People's Choice Best Interactive: After Dan Graham (David Han/Friend Generator) Best Immersive (Passive): 2nd Step (Joerg Courtial) Juried Technical Achievement: tx-reverse Excellence in Experience Design: Battlescar (Nico Casavecchia, Martin Allais) Excellence in Sound Design: Unheard (Zhechuan Zhang) Excellence in Visual Design: Ex Anima (Pierre Zandrowicz) Impact Award: State Power (Jeff Stanzler) Grand Jury Prize: The Industry (Mirka Duijn) === 2020 === People's Choice Best Interactive: Gravity VR (Fabito Rychter, Amir Admoni) Best Immersive (Passive): Warsaw Rising (Tomasz Dobosz) Juried Technical Achievement: The Cosmic Laughter of Cucci Binaca (Jonathan Sims) Excellence in Experience Design: Sleeping Eyes (Sojung Bahng, Sungeun Lee) Excellence in Sound Design: Symphony of Noise VR (Michaela Pnacekova) Excellence in Visual Design: Hominidae (Brian Andrews) Impact Award: Indirect Actions (Maranatha Hay) Grand Jury Prize: Minimum Mass (Raqi Syed, Areito Echevarria) === 2021 === FIVARS in FEB – People's Choice Best Interactive: CLAWS (created by Evan Neiden; directed by John Ertman) Best Immersive (Passive): Inside COVID 19 (Gary Yost, Adam Loften) FIVARS in FALL – People's Choice Best Interactive: Samsara (director: Hsin-Chien Huang) Best Immersive (Passive): The Invasion of Normandy Omaha Beach (director: Uli Futschik) Juried Technical Achievement: Dark Threads (director: Jonathon Corbiere) Excellence in Experience Design: Andy's World (director: Liquan Liu) Excellence in Sound Design: Symphony (director: Igor Cortadellas) Excellence in Visual Design: Mind VR Exploration (director: Deng Zuyun) Outstanding Performance: Lori Kovachevich, Lena's Journey (director: Wes Evans) Impact Award: Om Devi: Sheroes Revolution (director: Claudio Casale) Grand Jury Prize: Montegelato (director: Davide Rapp) === 2022 === FIVARS in FEB – People's Choice Best Interactive: Severance Theory: Welcome to Respite (Lyndsie Scoggin, United States) Best Immersive (Passive): Beescapes (Alan Nguyen, Australia) FIVARS in FALL – People's Choice Best Interactive: Namuanki (Kevin Mack, United States) Best Immersive (Passive): Reimagined Vol. 1: Nyssa (Julie Cavaliere, United States) Juried (Whole Year) Technical Achievement: Namuanki (Kevin Mack, United States) Excellence in Experience Design: Unframed: Hand Puppets, Paul Klee (Martin Charrière, Switzerland) Excellence in Visual Design: The Last Dance (Toshiaki Hanzaki, Japan) Excellence in Sound Design: Kingdom of Plants with David Attenborough (Iona McEwan, UK and USA) Outstanding Performance: Ari Tarr, OffRail (Ari Tarr, United States) Impact Award: Tearless (Gina Kim, South Korea) Grand Jury Prize: Klaxon. My dear sweet Friend (Nikita Shokhov, United States) === 2023 === People's Choice Best Interactive: PULSAR Best Immersive (Passive): Behind the Dish Juried Technical Achievement: VFC Excellence in Experience Design: Broken Spectre Excellence in Visual Design: Night Creatures Excellence in Sound Design: VFC Outstanding Performance: Origins Impact Award: LOU Grand Jury Prize: Stay Alive, My Son === 2024 ==
Empowerment (artificial intelligence)
Empowerment in the field of artificial intelligence formalises and quantifies (via information theory) the potential an agent perceives that it has to influence its environment. An agent which follows an empowerment maximising policy, acts to maximise future options (typically up to some limited horizon). Empowerment can be used as a (pseudo) utility function that depends only on information gathered from the local environment to guide action, rather than seeking an externally imposed goal, thus is a form of intrinsic motivation. The empowerment formalism depends on a probabilistic model commonly used in artificial intelligence. An autonomous agent operates in the world by taking in sensory information and acting to change its state, or that of the environment, in a cycle of perceiving and acting known as the perception-action loop. Agent state and actions are modelled by random variables ( S : s ∈ S , A : a ∈ A {\displaystyle S:s\in {\mathcal {S}},A:a\in {\mathcal {A}}} ) and time ( t {\displaystyle t} ). The choice of action depends on the current state, and the future state depends on the choice of action, thus the perception-action loop unrolled in time forms a causal bayesian network. == Definition == Empowerment ( E {\displaystyle {\mathfrak {E}}} ) is defined as the channel capacity ( C {\displaystyle C} ) of the actuation channel of the agent, and is formalised as the maximal possible information flow between the actions of the agent and the effect of those actions some time later. Empowerment can be thought of as the future potential of the agent to affect its environment, as measured by its sensors. E := C ( A t ⟶ S t + 1 ) ≡ max p ( a t ) I ( A t ; S t + 1 ) {\displaystyle {\mathfrak {E}}:=C(A_{t}\longrightarrow S_{t+1})\equiv \max _{p(a_{t})}I(A_{t};S_{t+1})} In a discrete time model, Empowerment can be computed for a given number of cycles into the future, which is referred to in the literature as 'n-step' empowerment. E ( A t n ⟶ S t + n ) = max p ( a t , . . . , a t + n − 1 ) I ( A t , . . . , A t + n − 1 ; S t + n ) {\displaystyle {\mathfrak {E}}(A_{t}^{n}\longrightarrow S_{t+n})=\max _{p(a_{t},...,a_{t+n-1})}I(A_{t},...,A_{t+n-1};S_{t+n})} The unit of empowerment depends on the logarithm base. Base 2 is commonly used in which case the unit is bits. === Contextual Empowerment === In general the choice of action (action distribution) that maximises empowerment varies from state to state. Knowing the empowerment of an agent in a specific state is useful, for example to construct an empowerment maximising policy. State-specific empowerment can be found using the more general formalism for 'contextual empowerment'. C {\displaystyle C} is a random variable describing the context (e.g. state). E ( A t n ⟶ S t + n ∣ C ) = ∑ c ∈ C p ( c ) E ( A t n ⟶ S t + n ∣ C = c ) {\displaystyle {\mathfrak {E}}(A_{t}^{n}\longrightarrow S_{t+n}{\mid }C)=\sum _{c{\in }C}p(c){\mathfrak {E}}(A_{t}^{n}\longrightarrow S_{t+n}{\mid }C=c)} == Application == Empowerment maximisation can be used as a pseudo-utility function to enable agents to exhibit intelligent behaviour without requiring the definition of external goals, for example balancing a pole in a cart-pole balancing scenario where no indication of the task is provided to the agent. Empowerment has been applied in studies of collective behaviour and in continuous domains. As is the case with Bayesian methods in general, computation of empowerment becomes computationally expensive as the number of actions and time horizon extends, but approaches to improve efficiency have led to usage in real-time control. Empowerment has been used for intrinsically motivated reinforcement learning agents playing video games, and in the control of underwater vehicles.
Instance (computer science)
In computer science, an instance or token (from metalogic and metamathematics) is a specific occurrence of a software element that is based on a type definition. When created, an occurrence is said to have been instantiated, and both the creation process and the result of creation are called instantiation. == Examples == Chat AI instance In chat-based AI systems, an assistant can be invoked across many independent conversation sessions (often called a thread), each with its own message history. A specific execution of the assistant over that session may be represented as a run (an execution on a thread). Class instance In object-oriented programming, an object created from a class type. Each instance of a class shares the class-defined structure and behavior but has its own identity and state. Procedural instance In some contexts (including Simula), each procedure call can be viewed as an instance of that procedure—an activation with its own parameters and local variables. Computer instance In cloud computing and virtualization, an instance commonly refers to a provisioned virtual machine or virtual server with an allocated combination of compute, memory, network, and storage resources. Polygonal model In computer graphics, a model may be instanced so it can be drawn multiple times with different transforms and parameters, improving performance by reusing shared geometry data. Program instance In a POSIX-oriented operating system, a running process is an instance of a program. It can be instantiated via system calls such as fork() and exec(). Each executing process is an instance of a program it has been instantiated from.
Intelligent agent
In artificial intelligence, an intelligent agent is an entity that perceives its environment, takes actions autonomously to achieve goals, and may improve its performance through machine learning or by acquiring knowledge. AI textbooks define artificial intelligence as the "study and design of intelligent agents," emphasizing that goal-directed behavior is central to intelligence. A specialized subset of intelligent agents, agentic AI (also known as an AI agent or simply agent), expands this concept by proactively pursuing goals, making decisions, and taking actions over extended periods. Intelligent agents can range from simple to highly complex. A basic thermostat or control system is considered an intelligent agent, as is a human being, or any other system that meets the same criteria—such as a firm, a state, or a biome. Intelligent agents operate based on an objective function, which encapsulates their goals. They are designed to create and execute plans that maximize the expected value of this function upon completion. For example, a reinforcement learning agent has a reward function, which allows programmers to shape its desired behavior. Similarly, an evolutionary algorithm's behavior is guided by a fitness function. Intelligent agents in artificial intelligence are closely related to agents in economics, and versions of the intelligent agent paradigm are studied in cognitive science, ethics, and the philosophy of practical reason, as well as in many interdisciplinary socio-cognitive modeling and computer social simulations. Intelligent agents are often described schematically as abstract functional systems similar to computer programs . To distinguish theoretical models from real-world implementations, abstract descriptions of intelligent agents are called abstract intelligent agents. Intelligent agents are also closely related to software agents—autonomous computer programs that carry out tasks on behalf of users. They are also referred to using a term borrowed from economics: a "rational agent". == Intelligent agents as the foundation of AI == The concept of intelligent agents provides a foundational lens through which to define and understand artificial intelligence. For instance, the influential textbook Artificial Intelligence: A Modern Approach (Russell & Norvig) describes: Agent: Anything that perceives its environment (using sensors) and acts upon it (using actuators). E.g., a robot with cameras and wheels, or a software program that reads data and makes recommendations. Rational Agent: An agent that strives to achieve the best possible outcome based on its knowledge and past experiences. "Best" is defined by a performance measure – a way of evaluating how well the agent is doing. Artificial Intelligence (as a field): The study and creation of these rational agents. Other researchers and definitions build upon this foundation. Padgham & Winikoff emphasize that intelligent agents should react to changes in their environment in a timely way, proactively pursue goals, and be flexible and robust (able to handle unexpected situations). Some also suggest that ideal agents should be "rational" in the economic sense (making optimal choices) and capable of complex reasoning, like having beliefs, desires, and intentions (BDI model). Kaplan and Haenlein offer a similar definition, focusing on a system's ability to understand external data, learn from that data, and use what is learned to achieve goals through flexible adaptation. Defining AI in terms of intelligent agents offers several key advantages: Avoids Philosophical Debates: It sidesteps arguments about whether AI is "truly" intelligent or conscious, like those raised by the Turing test or Searle's Chinese Room. It focuses on behavior and goal achievement, not on replicating human thought. Objective Testing: It provides a clear, scientific way to evaluate AI systems. Researchers can compare different approaches by measuring how well they maximize a specific "goal function" (or objective function). This allows for direct comparison and combination of techniques. Interdisciplinary Communication: It creates a common language for AI researchers to collaborate with other fields like mathematical optimization and economics, which also use concepts like "goals" and "rational agents." == Objective function == An objective function (or goal function) specifies the goals of an intelligent agent. An agent is deemed more intelligent if it consistently selects actions that yield outcomes better aligned with its objective function. In effect, the objective function serves as a measure of success. The objective function may be: Simple: For example, in a game of Go, the objective function might assign a value of 1 for a win and 0 for a loss. Complex: It might require the agent to evaluate and learn from past actions, adapting its behavior based on patterns that have proven effective. The objective function encapsulates all of the goals the agent is designed to achieve. For rational agents, it also incorporates the trade-offs between potentially conflicting goals. For instance, a self-driving car's objective function might balance factors such as safety, speed, and passenger comfort. Different terms are used to describe this concept, depending on the context. These include: Utility function: Often used in economics and decision theory, representing the desirability of a state. Objective function: A general term used in optimization. Loss function: Typically used in machine learning, where the goal is to minimize the loss (error). Reward Function: Used in reinforcement learning. Fitness Function: Used in evolutionary systems. Goals, and therefore the objective function, can be: Explicitly defined: Programmed directly into the agent. Induced: Learned or evolved over time. In reinforcement learning, a "reward function" provides feedback, encouraging desired behaviors and discouraging undesirable ones. The agent learns to maximize its cumulative reward. In evolutionary systems, a "fitness function" determines which agents are more likely to reproduce. This is analogous to natural selection, where organisms evolve to maximize their chances of survival and reproduction. Some AI systems, such as nearest-neighbor, reason by analogy rather than being explicitly goal-driven. However, even these systems can have goals implicitly defined within their training data. Such systems can still be benchmarked by framing the non-goal system as one whose "goal" is to accomplish its narrow classification task. Systems not traditionally considered agents, like knowledge-representation systems, are sometimes included in the paradigm by framing them as agents with a goal of, for example, answering questions accurately. Here, the concept of an "action" is extended to encompass the "act" of providing an answer. As a further extension, mimicry-driven systems can be framed as agents optimizing a "goal function" based on how closely the agent mimics the desired behavior. In generative adversarial networks (GANs) of the 2010s, an "encoder"/"generator" component attempts to mimic and improvise human text composition. The generator tries to maximize a function representing how well it can fool an antagonistic "predictor"/"discriminator" component. While symbolic AI systems often use an explicit goal function, the paradigm also applies to neural networks and evolutionary computing. Reinforcement learning can generate intelligent agents that appear to act in ways intended to maximize a "reward function". Sometimes, instead of setting the reward function directly equal to the desired benchmark evaluation function, machine learning programmers use reward shaping to initially give the machine rewards for incremental progress. Yann LeCun stated in 2018, "Most of the learning algorithms that people have come up with essentially consist of minimizing some objective function." AlphaZero chess had a simple objective function: +1 point for each win, and -1 point for each loss. A self-driving car's objective function would be more complex. Evolutionary computing can evolve intelligent agents that appear to act in ways intended to maximize a "fitness function" influencing how many descendants each agent is allowed to leave. The mathematical formalism of AIXI was proposed as a maximally intelligent agent in this paradigm. However, AIXI is uncomputable. In the real world, an intelligent agent is constrained by finite time and hardware resources, and scientists compete to produce algorithms that achieve progressively higher scores on benchmark tests with existing hardware. == Agent function == An intelligent agent's behavior can be described mathematically by an agent function. This function determines what the agent does based on what it has seen. A percept refers to the agent's sensory inputs at a single point in time. For example, a self-driving car's percepts might include camera images, lidar data, GPS coordinates, and speed r
Three-factor learning
In neuroscience and machine learning, three-factor learning is the combination of Hebbian plasticity with a third modulatory factor to stabilise and enhance synaptic learning. This third factor can represent various signals such as reward, punishment, error, surprise, or novelty, often implemented through neuromodulators. == Description == Three-factor learning introduces the concept of eligibility traces, which flag synapses for potential modification pending the arrival of the third factor, and helps temporal credit assignement by bridging the gap between rapid neuronal firing and slower behavioral timescales, from which learning can be done. Biological basis for Three-factor learning rules have been supported by experimental evidence. This approach addresses the instability of classical Hebbian learning by minimizing autocorrelation and maximizing cross-correlation between inputs.
Software requirements
Software requirements for a system are the description of what the system should do, the service or services that it provides and the constraints on its operation. The IEEE Standard Glossary of Software Engineering Terminology defines a requirement as: A condition or capability needed by a user to solve a problem or achieve an objective A condition or capability that must be met or possessed by a system or system component to satisfy a contract, standard, specification, or other formally imposed document A documented representation of a condition or capability as in 1 or 2 The activities related to working with software requirements can broadly be broken down into elicitation, analysis, specification, and management. Note that the wording Software requirements is additionally used in software release notes to explain, which depending on software packages are required for a certain software to be built/installed/used. == Elicitation == Elicitation is the gathering and discovery of requirements from stakeholders and other sources. A variety of techniques can be used such as joint application design (JAD) sessions, interviews, document analysis, focus groups, etc. Elicitation is the first step of requirements development. == Analysis == Analysis is the logical breakdown that proceeds from elicitation. Analysis involves reaching a richer and more precise understanding of each requirement and representing sets of requirements in multiple, complementary ways. Requirements Triage or prioritization of requirements is another activity which often follows analysis. This relates to Agile software development in the planning phase, e.g. by Planning poker, however it might not be the same depending on the context and nature of the project and requirements or product/service that is being built. == Specification == Specification involves representing and storing the collected requirements knowledge in a persistent and well-organized fashion that facilitates effective communication and change management. Use cases, user stories, functional requirements, and visual analysis models are popular choices for requirements specification. == Validation == Validation involves techniques to confirm that the correct set of requirements has been specified to build a solution that satisfies the project's business objectives, and to detect and correct errors in the requirements before implementation. == Management == Requirements change during projects and there are often many of them. Management of this change becomes paramount to ensuring that the correct software is built for the stakeholders. == Tool support for Requirements Engineering == === Tools for Requirements Elicitation, Analysis and Validation === Taking into account that these activities may involve some artifacts such as observation reports (user observation), questionnaires (interviews, surveys and polls), use cases, user stories; activities such as requirement workshops (charrettes), brainstorming, mind mapping, role-playing; and even, prototyping; software products providing some or all of these capabilities can be used to help achieve these tasks. There is at least one author who advocates, explicitly, for mind mapping tools such as FreeMind; and, alternatively, for the use of specification by example tools such as Concordion. Additionally, the ideas and statements resulting from these activities may be gathered and organized with wikis and other collaboration tools such as Trello. The features actually implemented and standards compliance vary from product to product. === Tools for Requirements Specification === A Software requirements specification (SRS) document might be created using general-purpose software like a word processor or one of several specialized tools. Some of these tools can import, edit, export and publish SRS documents. It may help to make SRS documents while following a standardised structure and methodology, such as ISO/IEC/IEEE 29148:2018. Likewise, software may or not use some standard to import or export requirements (such as ReqIF) or not allow these exchanges at all. === Tools for Requirements Document Verification === Tools of this kind verify if there are any errors in a requirements document according to some expected structure or standard. === Tools for Requirements Comparison === Tools of this kind compare two requirement sets according to some expected document structure and standard. === Tools for Requirements Merge and Update === Tools of this kind allow the merging and update of requirement documents. === Tools for Requirements Traceability === Tools of this kind allow tracing requirements to other artifacts such as models and source code (forward traceability) or, to previous ones such as business rules and constraints (backwards traceability). === Tools for Model-Based Software or Systems Requirement Engineering === Model-based systems engineering (MBSE) is the formalised application of modelling to support system requirements, design, analysis, verification and validation activities beginning in the conceptual design phase and continuing throughout development and later lifecycle phases. It is also possible to take a model-based approach for some stages of the requirements engineering and, a more traditional one, for others. Very many combinations might be possible. The level of formality and complexity depends on the underlying methodology involved (for instance, i is much more formal than SysML and, even more formal than UML) === Tools for general Requirements Engineering === Tools in this category may provide some mix of the capabilities mentioned previously and others such as requirement configuration management and collaboration. The features actually implemented and standards compliance vary from product to product. There are even more capable or general tools that support other stages and activities. They are classified as ALM tools.
Pythia (machine learning)
Pythia is an ancient text restoration model that recovers missing characters from damaged text input using deep neural networks. It was created by Yannis Assael, Thea Sommerschield, and Jonathan Prag, researchers from Google DeepMind and the University of Oxford. To study the society and the history of ancient civilisations, ancient history relies on disciplines such as epigraphy, the study of ancient inscribed texts. Hundreds of thousands of these texts, known as inscriptions, have survived to our day, but are often damaged over the centuries. Illegible parts of the text must then be restored by specialists, called epigraphists, in order to extract meaningful information from the text and use it to expand our knowledge of the context in which the text was written. Pythia takes as input the damaged text, and is trained to return hypothesised restorations of ancient Greek inscriptions, working as an assistive aid for ancient historians. Its neural network architecture works at both the character- and word-level, thereby effectively handling long-term context information, and dealing efficiently with incomplete word representations. Pythia is applicable to any discipline dealing with ancient texts (philology, papyrology, codicology) and can work in any language (ancient or modern).