AI Analysis X Ray

AI Analysis X Ray — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

• 

  Vero (app)

  Vero (stylized as VERO) is a social media platform and mobile app company. Vero markets itself as a social network free from advertisements, data mining and algorithms. == History == The app was founded by French-Lebanese billionaire Ayman Hariri who is the son of former Lebanese prime minister Rafic Hariri. The name is taken from the Italian word for true. The app launched officially in 2015 as an alternative to Facebook and their popular photo-blogging app Instagram. Within weeks of its release the app surged in popularity although users expressed mixed reports with some feeling confused about how the app worked. Cosplayers were early to adopt the app as their photo-sharing platform of choice, favouring the app's pinch and zoom magnification feature over Instagram's zoom feature. Other creative communities soon followed, and the app became popular with niche groups of makeup artists, tattoo artists, and skateboarders. In March 2018, Vero's popularity surged, partly helped by an exodus from Facebook and Instagram following the Cambridge Analytica data scandal. In the wake of the scandal, Vero devised an advertising campaign aimed at defected Facebook and Instagram users, hoping the app's policies and privacy settings would assuage concerns over sharing personal information on the internet. Within the space of one week, the app went from being a small service, akin to Ello or Peach, to being the most downloaded app in eighteen countries. In December 2020, Vero released its most significant update to date, Vero 2.0 which introduced new features including voice and video calls, game and app posts and bookmarks, and refinements to the UI. In October 2021, Vero introduced their Desktop app (beta) with multiple post options and a re-sizable multi-column feed. == Concept and funding == Vero's content feed resembles Instagram's although users can share a wider variety of content and the app has a chronological content feed whereas Facebook and Instagram's feeds are algorithm based. Vero's business plan is also distinct from similar social media apps. Whereas its competitors such as Facebook or Instagram make money from in-app advertising revenue and the sale of user data, Vero's business plan was to invite the first one million users to use the app for free then charge any subsequent users a subscription fee. The app was entirely funded by its founder and generated additional revenues by charging affiliate fees when someone buys a product they find on Vero. == Awards == Vero was recognized at the 2021 Webbys, being named as an Honoree in the Best Visual Design - Aesthetic Category. == Controversies == === Privacy === Vero has faced some criticism over the wording of their manifesto, in particular, the statement "Vero only collects the data we believe is necessary to provide users with a great experience and to ensure the security of their accounts." Because this policy does not explicitly state that the app will not sell data on to third parties some users fear that the need to monetise the app through data might prove too tempting. Users have also complained about not being able to delete their accounts. While this was never the case, the option was hidden deep in the app's settings. === Russian involvement === Although Vero remains transparent about the app's Russian development team, they have been caught up in concerns about Russian interference on social media platforms. The app's founder Ayman Hariri was quick to dismiss the remarks as xenophobic and defend the nationality of his employees, stating in an interview with Time Magazine; "At the end of the day, where people are from is really not how anybody should judge anyone". === Criticism of the app's founder === Until 2013, Vero's founder Ayman Harari was deputy CEO and chairman of Saudi Oger, the Saudi Arabian construction company which collapsed in 2017, mired by controversies over the welfare and treatment of their employees. However, Hariri is quick to point out that he divested from the firm in 2014 and the worker's rights violations occurred after he had left the company.

  Read more →
• 

  Forrest N. Iandola

  Forrest N. Iandola is an American computer scientist specializing in efficient AI. == Career == Iandola earned a PhD in Electrical Engineering and Computer Science from UC Berkeley in 2016, advised by Kurt Keutzer. As part of his dissertation, he co-authored SqueezeNet, a deep neural network for image classification optimized for smartphones and other mobile devices. Iandola and Keutzer went on to co-found DeepScale. The firm squeezes deep neural networks onto low-cost automotive-grade processors for use in driver assistance systems. Tesla acquired DeepScale in 2019. In 2020, he co-authored SqueezeBERT, an efficient neural network for natural language processing. In 2022, he joined Meta as an AI research scientist. His research at Meta includes developing efficient AI models, such as EfficientSAM and MobileLLM.

  Read more →
• 

  Noisy channel model

  The noisy channel model is a framework used in spell checkers, question answering, speech recognition, and machine translation. In this model, the goal is to find the intended word given a word where the letters have been scrambled in some manner. == In spell-checking == See Chapter B of. Given an alphabet Σ {\displaystyle \Sigma } , let Σ ∗ {\displaystyle \Sigma ^{}} be the set of all finite strings over Σ {\displaystyle \Sigma } . Let the dictionary D {\displaystyle D} of valid words be some subset of Σ ∗ {\displaystyle \Sigma ^{}} , i.e., D ⊆ Σ ∗ {\displaystyle D\subseteq \Sigma ^{}} . The noisy channel is the matrix Γ w s = Pr ( s | w ) {\displaystyle \Gamma _{ws}=\Pr(s|w)} , where w ∈ D {\displaystyle w\in D} is the intended word and s ∈ Σ ∗ {\displaystyle s\in \Sigma ^{}} is the scrambled word that was actually received. The goal of the noisy channel model is to find the intended word given the scrambled word that was received. The decision function σ : Σ ∗ → D {\displaystyle \sigma :\Sigma ^{}\to D} is a function that, given a scrambled word, returns the intended word. Methods of constructing a decision function include the maximum likelihood rule, the maximum a posteriori rule, and the minimum distance rule. In some cases, it may be better to accept the scrambled word as the intended word rather than attempt to find an intended word in the dictionary. For example, the word schönfinkeling may not be in the dictionary, but might in fact be the intended word. === Example === Consider the English alphabet Σ = { a , b , c , . . . , y , z , A , B , . . . , Z , . . . } {\displaystyle \Sigma =\{a,b,c,...,y,z,A,B,...,Z,...\}} . Some subset D ⊆ Σ ∗ {\displaystyle D\subseteq \Sigma ^{}} makes up the dictionary of valid English words. There are several mistakes that may occur while typing, including: Missing letters, e.g., leter instead of letter Accidental letter additions, e.g., misstake instead of mistake Swapping letters, e.g., recieved instead of received Replacing letters, e.g., fimite instead of finite To construct the noisy channel matrix Γ {\displaystyle \Gamma } , we must consider the probability of each mistake, given the intended word ( Pr ( s | w ) {\displaystyle \Pr(s|w)} for all w ∈ D {\displaystyle w\in D} and s ∈ Σ ∗ {\displaystyle s\in \Sigma ^{}} ). These probabilities may be gathered, for example, by considering the Damerau–Levenshtein distance between s {\displaystyle s} and w {\displaystyle w} or by comparing the draft of an essay with one that has been manually edited for spelling. == In machine translation == One naturally wonders if the problem of translation could conceivably be treated as a problem in cryptography. When I look at an article in Russian, I say: 'This is really written in English, but it has been coded in some strange symbols. I will now proceed to decode. See chapter 1, and chapter 25 of. Suppose we want to translate a foreign language to English, we could model P ( E | F ) {\displaystyle P(E|F)} directly: the probability that we have English sentence E given foreign sentence F, then we pick the most likely one E ^ = arg ⁡ max E P ( E | F ) {\displaystyle {\hat {E}}=\arg \max _{E}P(E|F)} . However, by Bayes law, we have the equivalent equation: E ^ = argmax E ∈ English P ( F ∣ E ) ⏞ translation model P ( E ) ⏞ language model {\displaystyle {\hat {E}}={\underset {E\in {\text{ English }}}{\operatorname {argmax} }}\overbrace {P(F\mid E)} ^{\text{translation model }}\overbrace {P(E)} ^{\text{language model}}} The benefit of the noisy-channel model is in terms of data: If collecting a parallel corpus is costly, then we would have only a small parallel corpus, so we can only train a moderately good English-to-foreign translation model, and a moderately good foreign-to-English translation model. However, we can collect a large corpus in the foreign language only, and a large corpus in the English language only, to train two good language models. Combining these four models, we immediately get a good English-to-foreign translator and a good foreign-to-English translator. The cost of noisy-channel model is that using Bayesian inference is more costly than using a translation model directly. Instead of reading out the most likely translation by arg ⁡ max E P ( E | F ) {\displaystyle \arg \max _{E}P(E|F)} , it would have to read out predictions by both the translation model and the language model, multiply them, and search for the highest number. == In speech recognition == Speech recognition can be thought of as translating from a sound-language to a text-language. Consequently, we have T ^ = argmax T ∈ Text P ( S ∣ T ) ⏞ speech model P ( T ) ⏞ language model {\displaystyle {\hat {T}}={\underset {T\in {\text{ Text }}}{\operatorname {argmax} }}\overbrace {P(S\mid T)} ^{\text{speech model }}\overbrace {P(T)} ^{\text{language model}}} where P ( S | T ) {\displaystyle P(S|T)} is the probability that a speech sound S is produced if the speaker is intending to say text T. Intuitively, this equation states that the most likely text is a text that's both a likely text in the language, and produces the speech sound with high probability. The utility of the noisy-channel model is not in capacity. Theoretically, any noisy-channel model can be replicated by a direct P ( T | S ) {\displaystyle P(T|S)} model. However, the noisy-channel model factors the model into two parts which are appropriate for the situation, and consequently it is generally more well-behaved. When a human speaks, it does not produce the sound directly, but first produces the text it wants to speak in the language centers of the brain, then the text is translated into sound by the motor cortex, vocal cords, and other parts of the body. The noisy-channel model matches this model of the human, and so it is appropriate. This is justified in the practical success of noisy-channel model in speech recognition. === Example === Consider the sound-language sentence (written in IPA for English) S = aɪ wʊd laɪk wʌn tuː. There are three possible texts T 1 , T 2 , T 3 {\displaystyle T_{1},T_{2},T_{3}} : T 1 = {\displaystyle T_{1}=} I would like one to. T 2 = {\displaystyle T_{2}=} I would like one too. T 3 = {\displaystyle T_{3}=} I would like one two. that are equally likely, in the sense that P ( S | T 1 ) = P ( S | T 2 ) = P ( S | T 3 ) {\displaystyle P(S|T_{1})=P(S|T_{2})=P(S|T_{3})} . With a good English language model, we would have P ( T 2 ) > P ( T 1 ) > P ( T 3 ) {\displaystyle P(T_{2})>P(T_{1})>P(T_{3})} , since the second sentence is grammatical, the first is not quite, but close to a grammatical one (such as "I would like one to [go]."), while the third one is far from grammatical. Consequently, the noisy-channel model would output T 2 {\displaystyle T_{2}} as the best transcription.

  Read more →
• 

  Noam Slonim

  Noam Slonim (Hebrew: נעם סלונים; born in Jerusalem) is an Israeli computer scientist, specializing in Natural Language Processing and the application of Large language models. He is a Research Scientist at Google Research Israel (since September 2025) and formerly an IBM Distinguished Engineer. He founded and served as Principal Investigator of Project Debater and led Language Model Utilization at IBM Research. Beyond his scientific achievements, Slonim had a writing and media career. He was a writer for Season 4 of The Cameric Five TV comedy show, published a weekly column in Haaretz on brain science, and co-created and wrote the Israeli sitcom Puzzle. He was also the head writer for Seasons 2 and 3 of the sitcom Ha-movilim and featured in the 2020 documentary The Debater. In October 2025, his debut novel, Questionable Memories, was published by Kinneret Publishing Group. == Education and research interests == Slonim graduated from the Hebrew University of Jerusalem in 1996 with a B.S. degree in Computer Science, Physics, and Mathematics. In 2002 he completed Ph.D. summa cum laude at the Interdisciplinary Center for Neural Computation at the Hebrew University, under the supervision of Professor Naftali Tishby. His thesis focused on the theory and applications of the Information Bottleneck method. From 2003 till 2006 he did post-doctoral studies at the Lewis-Sigler Institute for Integrative Genomics at Princeton University, working with Professor Bill Bialek and Professor Saeed Tavazoie. He joined IBM Research in 2007. Slonim holds over 30 patents (granted or pending) and has co-authored more than 100 scientific publications. In 2025, he joined Google Research Israel as a research scientist. == Research activities == From 1998 to 2003 he worked on the theory and applications of the Information Bottleneck method, suggesting various cluster analysis algorithms inspired by this method, and demonstrating the practical value of these algorithms on various domains. From 2003 to 2006 he worked on developing Machine Learning algorithms that rely on Information Theory concepts, and applied these algorithms to the analysis of various types of Genomics data. In 2011 he proposed to develop the first Artificial Intelligence system that can meaningfully participate in a full live debate with an expert human debater. This work gave rise to Project Debater, that debated expert human debaters in several live events during 2018 and 2019. In 2020, Slonim delivered the opening keynote at the EMNLP conference, describing the IBM Research work on developing Project Debater. From 2022 to 2025, he led IBM Research efforts applying large language models to practical use cases; in 2025 he moved to Google Research Israel as a Research Scientist. == Writing and video career == In 1996 Slonim was a writer for Season 4 of The Cameric Five TV comedy show. In 1997–1998 he published a weekly column in Haaretz newspaper, focused on brain science research. In 1997–1999 he co-created and co-wrote the Israeli sitcom, Puzzle. In 2008–2010 he was the head writer of Season 2 and Season 3 of the Israeli Sitcom, Ha-movilim. In 2020 he was featured in the documentary The Debater, an official selection of the 2020 Copenhagen International Documentary Film Festival. In 2025, his debut novel, Questionable Memories, was published by Kinneret Publishing Group.

  Read more →
• 

  Hardware for artificial intelligence

  Specialized computer hardware is often used to execute artificial intelligence (AI) programs faster, and with less energy, such as Lisp machines, neuromorphic engineering, event cameras, and physical neural networks. Since 2017, several consumer grade CPUs and SoCs have on-die NPUs. As of 2023, the market for AI hardware is dominated by GPUs. As of the 2020s, AI computation is dominated by graphics processing units (GPUs) and newer domain-specific accelerators such as Google's Tensor Processing Units (TPUs), AMD's Instinct MI300 series, and various on-device neural-processing units (NPUs) found in consumer hardware. == Scope == For the purposes of this article, AI hardware refers to computing components and systems specifically designed or optimized to accelerate artificial-intelligence workloads such as machine-learning training or inference. This includes general-purpose accelerators used for AI (for example, GPUs) and domain-specific accelerators (for example, TPUs, NPUs, and other AI ASICs). Event-based cameras are sometimes discussed in the context of neuromorphic computing, but they are input sensors rather than AI compute devices. Conversely, components such as memristors are basic circuit elements rather than specialized AI hardware when considered alone. == Lisp machines == Lisp machines were developed in the late 1970s and early 1980s to make artificial intelligence programs written in the programming language Lisp run faster. == Dataflow architecture == Dataflow architecture processors used for AI serve various purposes with varied implementations like the polymorphic dataflow Convolution Engine by Kinara (formerly Deep Vision), structure-driven dataflow by Hailo, and dataflow scheduling by Cerebras. == Component hardware == === AI accelerators === Since the 2010s, advances in computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer. By 2019, graphics processing units (GPUs), often with AI-specific enhancements, had displaced central processing units (CPUs) as the dominant means to train large-scale commercial cloud AI. OpenAI estimated the hardware compute used in the largest deep learning projects from Alex Net (2012) to Alpha Zero (2017), and found a 300,000-fold increase in the amount of compute needed, with a doubling-time trend of 3.4 months. === General-purpose GPUs for AI === Since the 2010s, graphics processing units (GPUs) have been widely used to train and deploy deep learning models because of their highly parallel architecture and high memory bandwidth. Modern data-center GPUs include dedicated tensor or matrix-math units that accelerate neural-network operations. In 2022, NVIDIA introduced the Hopper-generation H100 GPU, adding FP8 precision support and faster interconnects for large-scale model training. AMD and other vendors have also developed GPUs and accelerators aimed at AI and high-performance computing workloads. === Domain-specific accelerators (ASICs / NPUs) === Beyond general-purpose GPUs, several companies have developed application-specific integrated circuits (ASICs) and neural processing units (NPUs) tailored for AI workloads. Google introduced the Tensor Processing Unit (TPU) in 2016 for deep-learning inference, with later generations supporting large-scale training through dense systolic-array designs and optical interconnects. Other vendors have released similar devices—such as Apple's Neural Engine and various on-device NPUs—that emphasize energy-efficient inference in mobile or edge computing environments. === Memory and interconnects === AI accelerators rely on fast memory and inter-chip links to manage the large data volumes of training and inference. High-bandwidth memory (HBM) stacks, standardized as HBM3 in 2022, provide terabytes-per-second throughput on modern GPUs and ASICs. These accelerators are often connected through dedicated fabrics such as NVIDIA's NVLink and NVSwitch or optical interconnects used in TPU systems to scale performance across thousands of chips.

  Read more →
• 

  AI Text-to-image Tools: Free vs Paid (2026)

  Shopping for the best AI text-to-image tool? An AI text-to-image tool is software that uses machine learning to help you get more done — it keeps getting smarter as the underlying models improve. Pricing, accuracy, and the size of the model behind the tool are the three factors that most affect daily usefulness. Whether you are a beginner or a pro, the right AI text-to-image tool slots into your workflow and pays for itself fast. We tested the leading options and ranked them by quality, value, and ease of use.

  Read more →
• 

  Regular language

  In theoretical computer science and formal language theory, a regular language (also called a rational language) is a formal language that can be defined by a regular expression, in the strict sense in theoretical computer science (as opposed to many modern regular expression engines, which are augmented with features that allow the recognition of non-regular languages). Alternatively, a regular language can be defined as a language recognised by a finite automaton. The equivalence of regular expressions and finite automata is known as Kleene's theorem (after American mathematician Stephen Cole Kleene). In the Chomsky hierarchy, regular languages are the languages generated by Type-3 grammars. == Formal definition == The collection of regular languages over an alphabet Σ is defined recursively as follows: The empty language ∅ is a regular language. For each a ∈ Σ (a belongs to Σ), the singleton language {a} is a regular language. If A is a regular language, A (Kleene star) is a regular language. Due to this, the empty string language {ε} is also regular. If A and B are regular languages, then A ∪ B (union) and A • B (concatenation) are regular languages. No other languages over Σ are regular. See Regular expression § Formal language theory for syntax and semantics of regular expressions. == Examples == All finite languages are regular; in particular the empty string language {ε} = ∅ is regular. Other typical examples include the language consisting of all strings over the alphabet {a, b} which contain an even number of as, or the language consisting of all strings of the form: several as followed by several bs. A simple example of a language that is not regular is the set of strings {anbn | n ≥ 0}. Intuitively, it cannot be recognized with a finite automaton, since a finite automaton has finite memory and it cannot remember the exact number of a's. Techniques to prove this fact rigorously are given below. == Equivalent formalisms == A regular language satisfies the following equivalent properties: it is the language of a regular expression (by the above definition) it is the language accepted by a nondeterministic finite automaton (NFA) it is the language accepted by a deterministic finite automaton (DFA) it can be generated by a regular grammar it is the language accepted by an alternating finite automaton it is the language accepted by a two-way finite automaton it can be generated by a prefix grammar it can be accepted by a read-only Turing machine it can be defined in monadic second-order logic (Büchi–Elgot–Trakhtenbrot theorem) it is recognized by some finite syntactic monoid M, meaning it is the preimage {w ∈ Σ | f(w) ∈ S} of a subset S of a finite monoid M under a monoid homomorphism f : Σ → M from the free monoid on its alphabet the number of equivalence classes of its syntactic congruence is finite. (This number equals the number of states of the minimal deterministic finite automaton accepting L.) Properties 10. and 11. are purely algebraic approaches to define regular languages; a similar set of statements can be formulated for a monoid M ⊆ Σ. In this case, equivalence over M leads to the concept of a recognizable language. Some authors use one of the above properties different from "1." as an alternative definition of regular languages. Some of the equivalences above, particularly those among the first four formalisms, are called Kleene's theorem in textbooks. Precisely which one (or which subset) is called such varies between authors. One textbook calls the equivalence of regular expressions and NFAs ("1." and "2." above) "Kleene's theorem". Another textbook calls the equivalence of regular expressions and DFAs ("1." and "3." above) "Kleene's theorem". Two other textbooks first prove the expressive equivalence of NFAs and DFAs ("2." and "3.") and then state "Kleene's theorem" as the equivalence between regular expressions and finite automata (the latter said to describe "recognizable languages"). A linguistically oriented text first equates regular grammars ("4." above) with DFAs and NFAs, calls the languages generated by (any of) these "regular", after which it introduces regular expressions which it terms to describe "rational languages", and finally states "Kleene's theorem" as the coincidence of regular and rational languages. Other authors simply define "rational expression" and "regular expressions" as synonymous and do the same with "rational languages" and "regular languages". Apparently, the term regular originates from a 1951 technical report where Kleene introduced regular events and explicitly welcomed "any suggestions as to a more descriptive term". Noam Chomsky, in his 1959 seminal article, used the term regular in a different meaning at first (referring to what is called Chomsky normal form today), but noticed that his finite state languages were equivalent to Kleene's regular events. == Closure properties == The regular languages are closed under various operations, that is, if the languages K and L are regular, so is the result of the following operations: the set-theoretic Boolean operations: union K ∪ L, intersection K ∩ L, and complement L, hence also relative complement K − L. the regular operations: K ∪ L, concatenation ⁠ K ∘ L {\displaystyle K\circ L} ⁠, and Kleene star L. the trio operations: string homomorphism, inverse string homomorphism, and intersection with regular languages. As a consequence they are closed under arbitrary finite state transductions, like quotient K / L with a regular language. Even more, regular languages are closed under quotients with arbitrary languages: If L is regular then L / K is regular for any K. the reverse (or mirror image) LR. Given a nondeterministic finite automaton to recognize L, an automaton for LR can be obtained by reversing all transitions and interchanging starting and finishing states. This may result in multiple starting states; ε-transitions can be used to join them. == Decidability properties == Given two deterministic finite automata A and B, it is decidable whether they accept the same language. As a consequence, using the above closure properties, the following problems are also decidable for arbitrarily given deterministic finite automata A and B, with accepted languages LA and LB, respectively: Containment: is LA ⊆ LB ? Disjointness: is LA ∩ LB = {} ? Emptiness: is LA = {} ? Universality: is LA = Σ ? Membership: given a ∈ Σ, is a ∈ LB ? For regular expressions, the universality problem is NP-complete already for a singleton alphabet. For larger alphabets, that problem is PSPACE-complete. If regular expressions are extended to allow also a squaring operator, with "A2" denoting the same as "AA", still just regular languages can be described, but the universality problem has an exponential space lower bound, and is in fact complete for exponential space with respect to polynomial-time reduction. For a fixed finite alphabet, the theory of the set of all languages – together with strings, membership of a string in a language, and for each character, a function to append the character to a string (and no other operations) – is decidable, and its minimal elementary substructure consists precisely of regular languages. For a binary alphabet, the theory is called S2S. == Complexity results == In computational complexity theory, the complexity class of all regular languages is sometimes referred to as REGULAR or REG and equals DSPACE(O(1)), the decision problems that can be solved in constant space (the space used is independent of the input size). REGULAR ≠ AC0, since it (trivially) contains the parity problem of determining whether the number of 1 bits in the input is even or odd and this problem is not in AC0. On the other hand, REGULAR does not contain AC0, because the nonregular language of palindromes, or the nonregular language { 0 n 1 n : n ∈ N } {\displaystyle \{0^{n}1^{n}:n\in \mathbb {N} \}} can both be recognized in AC0. If a language is not regular, it requires a machine with at least Ω(log log n) space to recognize (where n is the input size). In other words, DSPACE(o(log log n)) equals the class of regular languages. In practice, most nonregular problems are studied in a setting with at least logarithmic space, as this is the amount of space required to store a pointer into the input tape. == Location in the Chomsky hierarchy == To locate the regular languages in the Chomsky hierarchy, one notices that every regular language is context-free. The converse is not true: for example, the language consisting of all strings having the same number of as as bs is context-free but not regular. To prove that a language is not regular, one often uses the Myhill–Nerode theorem and the pumping lemma. Other approaches include using the closure properties of regular languages or quantifying Kolmogorov complexity. Important subclasses of regular languages include: Finite languages, those containing only a finite number of words. These are regular la

  Read more →
• 

  Eurotra

  Eurotra was a machine translation project established and funded by the European Commission from 1978 until 1992. == History == In 1976, the European Commission started using the commercially developed machine translation system SYSTRAN with a plan to make it work for further languages than originally developed for (Russian-English and English-French), which however turned out to be difficult. This and the potential in existing systems within European research center, led to the decision in 1978 to start the project Eurotra, first through a preparatory Eurotra Coordination Group. Four years later, the European Commission and coordination group gained the approval of the European Parliament. The goal of the project as to create machine translation system for the official languages of the European Community, which at the time were Danish, Dutch, German, English, French, Italian, later including Greek, Spanish and Portuguese. However, as time passed, expectations became tempered; "Fully Automatic High Quality Translation" was not a reasonably attainable goal. The true character of Eurotra was eventually acknowledged to be in fact pre-competitive research rather than prototype development. The project was motivated by one of the founding principles of the EU: that all citizens had the right to read any and all proceedings of the Commission in their own language. As more countries joined, this produced a combinatorial explosion in the number of language pairs involved, and the need to translate every paper, speech and even set of meeting minutes produced by the EU into the other eight languages meant that translation rapidly became the overwhelming component in the administrative budget. To solve this problem Eurotra was devised. The project was unusual in that rather than consisting of a single research team, it had member groups distributed around the member countries, organised along language rather than national lines (for example, groups in Leuven and Utrecht worked closely together), and the secretariat was based at the European Commission in Luxembourg. The actual design of the project was unusual as MT projects go. Older systems, such as SYSTRAN, were heavily dictionary-based, with minor support for rearranging word order. More recent systems have often worked on a probabilistic approach, based on parallel corpora. Eurotra addressed the constituent structure of the text to be translated, going through first a syntactic parse followed by a second parse to produce a dependency structure followed by a final parse with a third grammar to produce what was referred to internally as Intermediate Representation (IR). Since all three modules were implemented as Prolog programs, it would then in principle be possible to put this structure backwards through the corresponding modules for another language to produce a translated text in any of the other languages. However, in practice this was not in fact how language pairs were implemented. The first "live" translation occupied a 4Mb Microvax running Ultrix and C-Prolog for a complete weekend some time in early 1987. The sentence, translated from English into Danish, was "Japan makes computers". The main problem faced by the system was the generation of so-called "Parse Forests" - often a large number of different grammar rules could be applied to any particular phrase, producing hundreds, even thousands of (often identical) parse trees. This used up huge quantities of computer store, slowing the whole process down unnecessarily. While Eurotra never delivered a "working" MT system, the project made a far-reaching long-term impact on the nascent language industries in European member states, in particular among the southern countries of Greece, Italy, Spain, and Portugal. There is at least one commercial MT system (developed by an academic/commercial consortium in Denmark) derived from Eurotra technology.

  Read more →
• 

  Ontology learning

  Ontology learning (ontology extraction, ontology augmentation generation, ontology generation, or ontology acquisition) is the automatic or semi-automatic creation of ontologies, including extracting the corresponding domain's terms and the relationships between the concepts that these terms represent from a corpus of natural language text, and encoding them with an ontology language for easy retrieval. As building ontologies manually is extremely labor-intensive and time-consuming, there is great motivation to automate the process. Typically, the process starts by extracting terms and concepts or noun phrases from plain text using linguistic processors such as part-of-speech tagging and phrase chunking. Then statistical or symbolic techniques are used to extract relation signatures, often based on pattern-based or definition-based hypernym extraction techniques. == Procedure == Ontology learning (OL) is used to (semi-)automatically extract whole ontologies from natural language text. The process is usually split into the following eight tasks, which are not all necessarily applied in every ontology learning system. === Domain terminology extraction === During the domain terminology extraction step, domain-specific terms are extracted, which are used in the following step (concept discovery) to derive concepts. Relevant terms can be determined, e.g., by calculation of the TF/IDF values or by application of the C-value / NC-value method. The resulting list of terms has to be filtered by a domain expert. In the subsequent step, similarly to coreference resolution in information extraction, the OL system determines synonyms, because they share the same meaning and therefore correspond to the same concept. The most common methods therefore are clustering and the application of statistical similarity measures. === Concept discovery === In the concept discovery step, terms are grouped to meaning bearing units, which correspond to an abstraction of the world and therefore to concepts. The grouped terms are these domain-specific terms and their synonyms, which were identified in the domain terminology extraction step. === Concept hierarchy derivation === In the concept hierarchy derivation step, the OL system tries to arrange the extracted concepts in a taxonomic structure. This is mostly achieved with unsupervised hierarchical clustering methods. Because the result of such methods is often noisy, a supervision step, e.g., user evaluation, is added. A further method for the derivation of a concept hierarchy exists in the usage of several patterns that should indicate a sub- or supersumption relationship. Patterns like “X, that is a Y” or “X is a Y” indicate that X is a subclass of Y. Such pattern can be analyzed efficiently, but they often occur too infrequently to extract enough sub- or supersumption relationships. Instead, bootstrapping methods are developed, which learn these patterns automatically and therefore ensure broader coverage. === Learning of non-taxonomic relations === In the learning of non-taxonomic relations step, relationships are extracted that do not express any sub- or supersumption. Such relationships are, e.g., works-for or located-in. There are two common approaches to solve this subtask. The first is based upon the extraction of anonymous associations, which are named appropriately in a second step. The second approach extracts verbs, which indicate a relationship between entities, represented by the surrounding words. The result of both approaches need to be evaluated by an ontologist to ensure accuracy. === Rule discovery === During rule discovery, axioms (formal description of concepts) are generated for the extracted concepts. This can be achieved, e.g., by analyzing the syntactic structure of a natural language definition and the application of transformation rules on the resulting dependency tree. The result of this process is a list of axioms, which, afterwards, is comprehended to a concept description. This output is then evaluated by an ontologist. === Ontology population === At this step, the ontology is augmented with instances of concepts and properties. For the augmentation with instances of concepts, methods based on the matching of lexico-syntactic patterns are used. Instances of properties are added through the application of bootstrapping methods, which collect relation tuples. === Concept hierarchy extension === In this step, the OL system tries to extend the taxonomic structure of an existing ontology with further concepts. This can be performed in a supervised manner with a trained classifier or in an unsupervised manner via the application of similarity measures. === Frame and Event detection === During frame/event detection, the OL system tries to extract complex relationships from text, e.g., who departed from where to what place and when. Approaches range from applying SVM with kernel methods to semantic role labeling (SRL) to deep semantic parsing techniques. == Tools == Dog4Dag (Dresden Ontology Generator for Directed Acyclic Graphs) is an ontology generation plugin for Protégé 4.1 and OBOEdit 2.1. It allows for term generation, sibling generation, definition generation, and relationship induction. Integrated into Protégé 4.1 and OBO-Edit 2.1, DOG4DAG allows ontology extension for all common ontology formats (e.g., OWL and OBO). Limited largely to EBI and Bio Portal lookup service extensions.

  Read more →
• 

  Bonnie Webber

  Bonnie Lynn Nash-Webber (born August 30, 1946) is a computational linguist. She is an honorary professor of intelligent systems in the Institute for Language, Cognition and Computation (ILCC) at the University of Edinburgh. == Education and career == Webber completed her PhD at Harvard University in 1978, advised by Bill Woods, while at the same time working with Woods at Bolt Beranek and Newman. == Career and research == Webber was appointed a professor at the University of Pennsylvania for 20 years before moving to Edinburgh in 1998. She has many academic descendants through her student at Pennsylvania, Martha E. Pollack. After retiring from the University of Edinburgh in 2016, she was listed by the university as an honorary professor. === Publications === Webber's doctoral dissertation, A Formal Approach to Discourse Anaphora, used formal logic to model the meanings of natural-language statements; it was published by Garland Publishers in 1979 in their Outstanding Dissertations in Linguistics Series. With Norman Badler and Cary Phillips, Webber is a co-author of the book Simulating Humans: Computer Graphics Animation and Control (Oxford University Press, 1993). With Aravind Joshi and Ivan Sag she is a co-editor of Elements of Discourse Understanding, with Nils Nilsson she is co-editor of Readings in Artificial Intelligence, and with Barbara Grosz and Karen Spärck Jones she is co-editor of Readings in Natural Language Processing. === Awards and honours === Webber was appointed a Founding Fellow of the Association for the Advancement of Artificial Intelligence (AAAI) in 1990, and was elected a Fellow of the Royal Society of Edinburgh (FRSE) in 2004. She served as president of the Association for Computational Linguistics (ACL) in 1980, and became a Fellow of the Association for Computational Linguistics in 2012, "for significant contributions to discourse structure and discourse-based interpretation". In 2020, she was awarded the Association for Computational Linguistics Lifetime Achievement Award.

  Read more →
• 

  Comparison of machine translation applications

  Machine translation is an algorithm which attempts to translate text or speech from one natural language to another. == General information == Basic general information for popular machine translation applications. == Languages features comparison == The following table compares the number of languages which the following machine translation programs can translate between. (Moses and Moses for Mere Mortals allow you to train translation models for any language pair, though collections of translated texts (parallel corpus) need to be provided by the user. The Moses site provides links to training corpora.) This is not an all-encompassing list. Some applications have many more language pairs than those listed below. This is a general comparison of key languages only. A full and accurate list of language pairs supported by each product should be found on each of the product's websites. === Multi-pair translations === === Paired translations ===

  Read more →
• 

  Trevor Hastie

  Trevor John Hastie (born 27 June 1953) is an American statistician and computer scientist. He is currently serving as the John A. Overdeck Professor of Mathematical Sciences and Professor of Statistics at Stanford University. Hastie is known for his contributions to applied statistics, especially in the field of machine learning, data mining, and bioinformatics. He has authored several popular books in statistical learning, including The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Hastie has been listed as an ISI Highly Cited Author in Mathematics by the ISI Web of Knowledge. He also contributed to the development of S. == Education and career == Hastie was born on 27 June 1953 in South Africa. He received his B.S. in statistics from the Rhodes University in 1976 and master's degree from University of Cape Town in 1979. Hastie joined the doctoral program at Stanford University in 1980 and received his Ph.D. in 1984 under the supervision of Werner Stuetzle. His dissertation was "Principal Curves and Surfaces". Hastie began his professional career in 1977 with the South African Medical Research Council. After receiving his master's degree in 1979, he spent a year interning at the London School of Hygiene & Tropical Medicine, the Johnson Space Center in Houston, and the Biomath department at Oxford University. After receiving his doctoral degree from Stanford, Hastie returned to South Africa to work with his former employer South African Medical Research Council. He returned to United States in 1986 and joined the AT&T Bell Laboratories in Murray Hill, New Jersey and remained there for nine years. Working with John Chambers, he co-directed the development of the S programming language. He joined Stanford University in 1994 as Associate Professor in Statistics and Biostatistics. He was promoted to full Professor in 1999. During the period 2006–2009, he was the chair of the Department of Statistics at Stanford University. In 2013 he was named the John A. Overdeck Professor of Mathematical Sciences. == Awards and honors == Hastie is a Fellow of the Royal Statistical Society since 1979. He is also an elected Fellow of several professional and scholarly societies, including the Institute of Mathematical Statistics, the American Statistical Association, and the South African Statistical Society. He is a recipient of 'Myrto Lefkopolou Distinguished Lectureship' award of Biostatistics Department at the Harvard School of Public Health. In 2018, he was elected a member of the National Academy of Sciences. In 2019 Hastie became a foreign member of the Royal Netherlands Academy of Arts and Sciences. Hastie was named for the C.R. and Bhargavi Rao Prize in 2025. Hastie and Hui Zou received the 2025 Founders of Statistics prize for their elastic net paper. == Publications == Hastie is a prolific author of scientific works on numerous topics in applied statistics, including statistical learning, data mining, statistical computing, and bioinformatics. He along with his collaborators has authored about 125 scientific articles. Many of Hastie's scientific articles were coauthored by his longtime collaborator, Robert Tibshirani. Hastie has been listed as an ISI Highly Cited Author in Mathematics by the ISI Web of Knowledge. He has coauthored the following books: T. Hastie and R. Tibshirani, Generalized Additive Models, Chapman and Hall, 1990. J. Chambers and T. Hastie, Statistical Models in S, Wadsworth/Brooks Cole, 1991. T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning: Prediction, Inference and Data Mining, Second Edition, Springer Verlag, 2009 (available for free from the author's website). G. James, D. Witten, T. Hastie, R. Tibshirani, An Introduction to Statistical Learning with Applications in R, Springer Verlag, 2013 (available for free from the co-author's website). T. Hastie, R. Tibshirani, M. Wainwright, Statistical Learning with Sparsity: the Lasso and Generalizations, CRC Press, 2015 (available for free from the author's website). Bradley Efron; Trevor Hastie (2016). Computer Age Statistical Inference. Cambridge University Press. ISBN 9781107149892.

  Read more →
• 

  MY F.C.

  MY F.C. is a freemium app designed to organise and administer football teams. It is developed by MY F.C. Limited, a private company headquartered in Auckland, New Zealand. The app allows users to build a team by adding players and from there they can create trainings and matches, keep up with relevant news in the curated newsfeed, record statistics both individually and team based, follow the games live in the match-centre. The app also features integrated lineup builder with custom team kits. == History == Founders Sam Jenkins, Mike Simpson and Sam Jasper started MY F.C. in 2015 to help them "run their football lives". The app was launched on Android and iOS on 14 February 2017. == Accolades == MY F.C. won the first place prize at Bank of New Zealand Start-up Alley 2017 competition that aims to discover New Zealand start-ups who are doing innovative work and ready to establish themselves as long-term, sustainable businesses. The prize package included $15,000 and a trip to San Francisco.

  Read more →
• 

  RAMnets

  RAMnets is one of the oldest practical neurally inspired classification algorithms. The RAMnets is also known as a type of "n-tuple recognition method" or "weightless neural network". == Algorithm == Consider (let us say N) sets of n distinct bit locations are selected randomly. These are the n-tuples. The restriction of a pattern to an n-tuple can be regarded as an n-bit number which, together with the identity of the n-tuple, constitutes a `feature' of the pattern. The standard n-tuple recognizer operates simply as follows: A pattern is classified as belonging to the class for which it has the most features in common with at least one training pattern of that class. This is the Θ {\displaystyle \Theta } = 0 case of a more general rule whereby the class assigned to unclassified pattern u is a c r g m a x ( ∑ i = 1 N Θ ( ∑ v ∈ D c δ ( α i ( u ) , α i ( v ) ) ) ) {\displaystyle {\begin{aligned}{\underset {c}{a}}rgmax(\sum _{i=1}^{N}\Theta (\sum _{v\in D_{c}}\delta (\alpha _{i}(u),\alpha _{i}(v))))\end{aligned}}} where Dc is the set of training patterns in class c, Θ ( x ) {\displaystyle \Theta (x)} = x for 0 ≤ x ≤ θ {\displaystyle 0\leq x\leq \theta } , Θ ( x ) = θ {\displaystyle \Theta (x)=\theta } for x ≥ θ {\displaystyle x\geq \theta } , δ i , j {\displaystyle \delta _{i,j}} is the Kronecker delta( δ i , j {\displaystyle \delta _{i,j}} =1 if i=j and 0 otherwise.)and ( α i ( u ) ) {\displaystyle (\alpha _{i}(u))} is the ith feature of the pattern u: ∑ j = 0 n − 1 u η i ( j ) 2 j {\displaystyle \sum _{j=0}^{n-1}u_{\eta }i(j)2^{j}} Here uk is the kth bit of u and u η i ( j ) {\displaystyle u_{\eta }i(j)} is the jth bit location of the ith n-tuple. With C classes to distinguish, the system can be implemented as a network of NC nodes, each of which is a random access memory (RAM); hence the term RAMnet. The memory content m c i α {\displaystyle m_{ci\alpha }} at address α {\displaystyle \alpha } of the ith node allocated to class c is set to m c i α {\displaystyle m_{ci\alpha }} = Θ ( ∑ v ∈ D c δ ( α , α i ( v ) ) ) {\displaystyle \Theta (\sum _{v\in D_{c}}\delta (\alpha ,\alpha _{i}(v)))} In the usual θ {\displaystyle \theta } = 1 case, the 1-bit content of m c i α {\displaystyle m_{ci\alpha }} is set if any pattern of Dc has feature α {\displaystyle \alpha } and unset otherwise. Recognition is accomplished by summing the contents of the nodes of each class at the addresses given by the features of the unclassified pattern. That is, pattern u is assigned to class a c r g m a x ( ∑ i = 1 N m c i α ( u ) ) {\displaystyle {\begin{aligned}{\underset {c}{a}}rgmax(\sum _{i=1}^{N}m_{ci\alpha }(u))\end{aligned}}} == RAM-discriminators and WiSARD == The RAMnets formed the basis of a commercial product known as WiSARD (Wilkie, Stonham and Aleksander Recognition Device) was the first artificial neural network machine to be patented. A RAM-discriminator consists of a set of X one-bit word RAMs with n inputs and a summing device (Σ). Any such RAM-discriminator can receive a binary pattern of X⋅n bits as input. The RAM input lines are connected to the input pattern by means of a biunivocal pseudo-random mapping. The summing device enables this network of RAMs to exhibit – just like other ANN models based on synaptic weights – generalization and noise tolerance. In order to train the discriminator one has to set all RAM memory locations to 0 and choose a training set formed by binary patterns of X⋅n bits. For each training pattern, a 1 is stored in the memory location of each RAM addressed by this input pattern. Once the training of patterns is completed, RAM memory contents will be set to a certain number of 0's and 1's. The information stored by the RAM during the training phase is used to deal with previous unseen patterns. When one of these is given as input, the RAM memory contents addressed by the input pattern are read and summed by Σ. The number r thus obtained, which is called the discriminator response, is equal to the number of RAMs that output 1. r reaches the maximum X if the input belongs to the training set. r is equal to 0 if no n-bit component of the input pattern appears in the training set (not a single RAM outputs 1). Intermediate values of r express a kind of “similarity measure” of the input pattern with respect to the patterns in the training set. A system formed by various RAM-discriminators is called WiSARD. Each RAM-discriminator is trained on a particular class of patterns, and classification by the multi-discriminator system is performed in the following way. When a pattern is given as input, each RAM-discriminator gives a response to that input. The various responses are evaluated by an algorithm which compares them and computes the relative confidence c of the highest response (e.g., the difference d between the highest response and the second highest response, divided by the highest response). A schematic representation of a RAM-discriminator and a 10 RAM-discriminator WiSARD is shown in Figure 1.

  Read more →
• 

  Node2vec

  node2vec is an algorithm to generate vector representations of nodes on a graph. The node2vec framework learns low-dimensional representations for nodes in a graph through the use of random walks through a graph starting at a target node. It is useful for a variety of machine learning applications. node2vec follows the intuition that random walks through a graph can be treated like sentences in a corpus. Each node in a graph is treated like an individual word, and a random walk is treated as a sentence. By feeding these "sentences" into a skip-gram, or by using the continuous bag of words model, paths found by random walks can be treated as sentences, and traditional data-mining techniques for documents can be used. The algorithm generalizes prior work which is based on rigid notions of network neighborhoods, and argues that the added flexibility in exploring neighborhoods is the key to learning richer representations of nodes in graphs. The algorithm is considered one of the best graph classifiers.

  Read more →