AI Writing Tools

Explore the best AI Writing Tools — independent reviews, comparisons, pricing and step-by-step how-to guides, curated by Aizhi.

Site-specific browser

A site-specific browser (SSB) is a software application dedicated to accessing pages from a single source (site) on a computer network such as the Internet or a private intranet. SSBs typically simplify the more complex functions of a web browser by excluding the menus, toolbars and browser graphical user interface associated with functions that are external to the workings of a single site. Modern site-specific browsers range from simple browser windows without navigation controls to sophisticated desktop applications built with frameworks like Electron that bundle entire browser engines. This evolution has enabled many popular desktop applications to be built using web technologies, effectively making them advanced site-specific browsers. == History == === Early development === One of the earliest examples of an SSB was MacDICT, a Mac OS 9 application that accessed various websites to define, translate, or find synonyms for words typed into a text box. However, the first general-purpose SSB is considered to be Bubbles, which launched in late 2005 on the Windows platform. Bubbles introduced the term "Site Specific Extensions" for SSB userscripts and created the first SSB JavaScript API. In 2007, Mozilla announced Prism (originally called WebRunner), a project to integrate web applications with the desktop. That same year, Todd Ditchendorf, a former Apple Dashboard engineer, released Fluid for macOS. On 2 September 2008, Google Chrome was released with a built-in "Create application shortcut" feature, bringing SSB functionality to mainstream users. This feature allowed any website to be launched in a separate window without the browser interface. === Modern era === The landscape of site-specific browsers changed dramatically with the introduction of Electron in 2013 (originally called Atom Shell). Electron combined Chromium and Node.js into a single runtime, enabling developers to build desktop applications using web technologies. This framework has since powered applications used by hundreds of millions of users, including Visual Studio Code, Slack, Discord, and Microsoft Teams. In 2015, the concept of Progressive Web Apps (PWAs) was introduced by Google engineers Alex Russell and Frances Berriman, representing a parallel evolution in web-to-desktop technology. While PWAs share similar goals with SSBs, they follow web standards and can be installed directly from browsers. More recently, alternative frameworks like Tauri have emerged, offering significantly smaller application sizes by using the system's native web renderer instead of bundling Chromium. == Technical implementation == Site-specific browsers can be implemented through various approaches: === Browser-based SSBs === The simplest form of SSB is created through browser features that allow websites to run in separate windows without the standard browser interface. Modern Chromium-based browsers offer "Install as app" or "Create shortcut" functionality that creates a dedicated window for a specific website. These SSBs share the browser's underlying engine and resources but operate in isolated windows. === Framework-based SSBs === More sophisticated SSBs are built using application frameworks: Electron: Bundles a complete Chromium browser with Node.js, resulting in applications of 85MB or larger. Each Electron application runs its own browser instance, providing full access to system APIs but consuming significant resources. Tauri: Uses the operating system's native web rendering engine (WebView2 on Windows, WebKit on macOS, and WebKitGTK on Linux), resulting in applications typically 2.5-10MB in size. Other frameworks: Include Neutralino.js (ultra-lightweight using system browser), Wails (Go-based), and the Chromium Embedded Framework (CEF). == Comparison with Progressive Web Apps == While site-specific browsers and Progressive Web Apps (PWAs) share the goal of bringing web content to the desktop, they differ in several key aspects: == Applications == Site-specific browsers have become the foundation for many popular desktop applications: Communication and collaboration: Many modern communication tools are built as SSBs, including Slack, Discord, Microsoft Teams, and WhatsApp Desktop. These applications benefit from web-based development while providing desktop integration. Development tools: Visual Studio Code, used by 73.6% of developers according to Stack Overflow's 2024 survey, is built with Electron, as are Atom and GitHub Desktop. Productivity software: Applications like Notion, Obsidian, and various project management tools use SSB technology to provide consistent experiences across platforms. Security and Privacy: Web browsers can be modified to only have access to a single site, in order to protect the security and privacy of the user via compartmentalization == Security and performance == === Memory usage === Framework-based SSBs, particularly those using Electron, are known for high memory consumption. Studies show Electron applications typically use 120-300MB at baseline, with complex applications consuming significantly more. This is approximately 5-10 times more memory than equivalent native applications. === Security considerations === SSBs can provide security benefits through process isolation, where each application runs in its own sandboxed environment. However, bundling an entire browser engine also means each application must be updated independently to patch security vulnerabilities. Research presented at the Network and Distributed System Security (NDSS) Symposium has identified various security challenges specific to Electron applications. === Bundle sizes === The choice of framework significantly impacts application size: Electron applications: 85MB+ (includes full Chromium) Tauri applications: 2.5-10MB (uses system WebView) Browser-based SSBs: No additional download (uses existing browser) == Software == === Browser support === Most modern browsers provide some form of SSB functionality: Chromium-based browsers (Google Chrome, Microsoft Edge, Brave, Opera, Vivaldi): "Install as app" or "Create shortcut" feature Safari: "Add to Dock" feature in macOS Sonoma (2023) Firefox: Removed SSB support in December 2020 (version 85) GNOME Web: "Install Site as Web Application" feature === Standalone tools === ==== Active ==== WebCatalog (Windows, macOS, Linux) – Manages multiple SSBs with isolated storage Fluid (macOS) – Pioneering SSB creator for Mac Unite (macOS) – Creates SSBs with customization options Coherence X (macOS) – Advanced SSB creation tool Pake (cross-platform) – Open-source SSB creator Wavebox (cross-platform) – Workspace browser with SSB features ==== Discontinued ==== Mozilla Prism – Cross-platform SSB creator (discontinued 2011) Nativefier – Command-line SSB creator (discontinued 2023) Epichrome – macOS SSB creator (discontinued 2021) === Development frameworks === Electron – Most popular framework, bundles Chromium and Node.js Tauri – Rust-based framework using system WebView Chromium Embedded Framework (CEF) – C++ library for embedding Chromium Neutralino.js – Lightweight framework using system browser Wails – Go-based framework for web frontends
Read more →
Is an AI Sales Assistant Worth It in 2026?

Shopping for the best AI sales assistant? An AI sales assistant 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 sales assistant slots into your workflow and pays for itself fast. Below we compare features, pricing, and real output so you can choose with confidence.
Read more →
Ziad Obermeyer

Ziad Obermeyer (Arabic: زياد أوبرماير) is a Lebanese American physician and researcher whose work focuses on machine learning, health policy, and clinical decision-making in medicine. He is the Blue Cross of California Distinguished Associate Professor at the UC Berkeley School of Public Health, a Chan Zuckerberg Biohub investigator, and a research associate at the National Bureau of Economic Research. He is known for his research on racial bias in health care algorithms and the use of artificial intelligence in health care. == Early life and education == Obermeyer was born in Beirut, Lebanon, and raised in Cambridge, Massachusetts. He earned a Bachelor of Arts degree from Harvard College and a Master of Philosophy (M.Phil.) in History and Science from the University of Cambridge. He received his Doctor of Medicine (M.D.) from Harvard Medical School in 2008. Before pursuing medicine, Obermeyer worked as a consultant at McKinsey & Company, advising pharmaceutical and global health clients in New Jersey, Geneva, and Tokyo. After completing his medical degree, he trained as an emergency physician at Mass General Brigham (MGB) in Boston, Massachusetts. He later continued practicing emergency medicine at the Fort Defiance Indian Hospital on the Navajo Nation in Arizona. == Academic career == Obermeyer served as an Assistant Professor at Harvard Medical School from 2014 to 2020. In 2020, he joined the University of California, Berkeley as an Associate Professor and the Blue Cross of California Distinguished Professor at the School of Public Health. == Research focus == === Algorithmic racial bias in healthcare === In 2019, Obermeyer and economist Sendhil Mullainathan examined a commercial healthcare algorithm by UnitedHealth Group, used in hospitals and by insurers to identify patients with complex health needs. The study found that the algorithm underestimated the health needs of Black patients compared to white patients with similar conditions and that reformulating it would reduce racial bias. In 2020, Obermeyer analyzed an algorithm used to allocate CARE Act relief funding to hospitals. The study identified allocation patterns that favored hospitals with higher revenues over hospitals serving larger numbers of COVID-19 patients who are predominantly Black. === Clinical decision-making === In 2021, Obermeyer and colleagues examined physician decision-making in cardiac care using machine learning models. The study found that physicians misdiagnose cases when they rely on symptoms representative of a heart attack, such as chest pain, over other symptoms. === Pain assessment === Obermeyer developed a deep learning approach to investigate the severity of osteoarthritis in underserved communities. == Policy and regulatory work == Following the publication of the 2019 algorithmic racial bias study, the New York Department of Financial Services and Department of Health launched an investigation into UnitedHealth Group's algorithm, requesting that the company cease using it, citing discriminatory business practices. Also related to this study, in December 2019, Democratic Senators Cory Booker and Ron Wyden released letters to the Federal Trade Commission and Centers for Medicare and Medicaid Services asking to investigate potential discrimination in decision-making algorithms against marginalized communities in healthcare. The senators also wrote to major healthcare companies, including Aetna and Blue Cross Blue Shield, about their internal safeguards against racial bias in their technology. In 2021, Obermeyer and colleagues at the University of Chicago Booth School of Business released the Algorithmic Bias Playbook, a resource for policymakers and technical teams working in healthcare on how to measure and mitigate algorithmic racial bias. Obermeyer testified before the U.S. Senate Financial Committee in February 2024 on artificial intelligence in healthcare, recommending transparency requirements for AI developers and independent algorithm evaluations. In December 2025, he testified before the United States House Committee on Oversight and Government Reform on the role of AI in affordable healthcare and the impact of its integration on the workforce. == Organizations == In 2021, Obermeyer cofounded Nightingale Open Science, a non-profit that creates new medical imaging datasets available for research, and Dandelion Health, a health data analytics company. In June 2023, the company launched a program to audit and evaluate the performance of algorithms to identify potential racial, ethnic, and geographic bias, funded by the Gordon and Betty Moore Foundation and the SCAN Foundation. Dandelion Health partnered with the American Heart Association in 2025 to power an AI assessment lab for cardiovascular algorithms. Obermeyer is a founding faculty member of the University of California, Berkeley–University of California, San Francisco joint program in computational precision health. == Recognition == TIME magazine named Obermeyer one of the 100 most influential people in artificial intelligence in 2023. He has served as a Chan Zuckerberg Biohub Investigator since 2022, and as a Research Associate at the National Bureau of Economic Research since 2023. He was designated an Emerging Leader by the National Academy of Medicine in 2020. Obermeyer's racial bias study received the Willard G. Manning Memorial Award for the Best Research in Health Econometrics from the American Society of Health Economists (ASHEcon) in 2021 and the Responsible Business Education Award from the Financial Times in 2022.
Read more →
Mark Heimann

Mark A. Heimann is an American chess grandmaster and machine learning researcher. == Chess career == Heimann began playing chess at the age of 5 after his father bought him and his twin brother Alexander a chess set. He then won several national grade-level championships as well as the Pennsylvania and Ohio state championships in middle school and high school. In October 2007, he was ranked as the national #2 under-14 player, only behind future grandmaster Marc Tyler Arnold. In the February 2008 national rankings, he moved up to being the top-ranked under-14 player. In December 2012, he played for Washington University St. Louis' "A" team in the Pan-American Intercollegiate Chess Championships, where he was the second-most successful player, recording 4 wins, 1 draw, and 1 loss. The university's team also won the Division II championship title. In three tournaments between September and December 2022, Heimann earned three international master title norms, earning the international master title at the age of 29. In November 2024, he scored a GM norm at the U.S. Masters Chess Championship. He finished the event in joint-6th place. The following week, at the Saint Louis Masters tournament, he earned his final grandmaster norm and crossed 2500 in live rating, achieving the Grandmaster title. It was formally awarded to him in April 2025. == Research career == He obtained a bachelor's degree from Washington University in St. Louis in the School of Arts and Sciences and got his PhD from the University of Michigan. He is a machine learning researcher at Lawrence Livermore National Laboratory. == Personal life == Outside of chess and research, he also plays several instruments and is a competitive powerlifter.
Read more →
Retained mode

Retained mode in computer graphics is a major pattern of API design in graphics libraries, in which the graphics library, instead of the client, retains the scene (complete object model of the rendering primitives) to be rendered and the client calls into the graphics library do not directly cause actual rendering, but make use of extensive indirection to resources, managed – thus retained – by the graphics library. It does not preclude the use of double-buffering. Immediate mode is an alternative approach. Historically, retained mode has been the dominant style in GUI libraries; however, both can coexist in the same library and are not necessarily exclusionary in practice. == Overview == In retained mode the client calls do not directly cause actual rendering, but instead update an abstract internal model (typically a list of objects) which is maintained within the library's data space. This allows the library to optimize when actual rendering takes place along with the processing of related objects. Some techniques to optimize rendering include: managing double buffering treatment of hidden surfaces by backface culling/occlusion culling (Z-buffering) only transferring data that has changed from one frame to the next from the application to the library Example of coexistence with immediate mode in the same library is OpenGL. OpenGL has immediate mode functions that can use previously defined server side objects (textures, vertex buffers and index buffers, shaders, etc.) without resending unchanged data. Examples of retained mode rendering systems include Windows Presentation Foundation, SceneKit on macOS, and PHIGS.
Read more →
How to Choose an AI Analytics Tool

Looking for the best AI analytics tool? An AI analytics tool is software that uses machine learning to help you get more done — it can save you hours every week by automating repetitive work. Most options offer a generous free tier, with paid plans unlocking higher limits, faster processing, and team features. Whether you are a beginner or a pro, the right AI analytics tool slots into your workflow and pays for itself fast. Read on for hands-on impressions, pricing tiers, and the standout features that matter.
Read more →
Powerset construction

In the theory of computation and automata theory, the powerset construction or subset construction is a standard method for converting a nondeterministic finite automaton (NFA) into a deterministic finite automaton (DFA) that recognizes the same formal language. It is important in theory because it establishes that NFAs, despite their additional flexibility, are unable to recognize any language that cannot be recognized by some DFA. It is also important in practice for converting easier-to-construct NFAs into more efficiently executable DFAs. However, if the NFA has n states, the resulting DFA may have up to 2n states, an exponentially larger number, which sometimes makes the construction impractical for large NFAs. The construction, sometimes called the Rabin–Scott powerset construction (or subset construction) to distinguish it from similar constructions for other types of automata, was first published by Michael O. Rabin and Dana Scott in 1959. == Intuition == To simulate the operation of a DFA on a given input string, one needs to keep track of a single state at any time: the state that the automaton will reach after seeing a prefix of the input. In contrast, to simulate an NFA, one needs to keep track of a set of states: all of the states that the automaton could reach after seeing the same prefix of the input, according to the nondeterministic choices made by the automaton. If, after a certain prefix of the input, a set S of states can be reached, then after the next input symbol x the set of reachable states is a deterministic function of S and x. Therefore, the sets of reachable NFA states play the same role in the NFA simulation as single DFA states play in the DFA simulation, and in fact the sets of NFA states appearing in this simulation may be re-interpreted as being states of a DFA. == Construction == The powerset construction applies most directly to an NFA that does not allow state transformations without consuming input symbols (aka: "ε-moves"). Such an automaton may be defined as a 5-tuple (Q, Σ, T, q0, F), in which Q is the set of states, Σ is the set of input symbols, T is the transition function (mapping a state and an input symbol to a set of states), q0 is the initial state, and F is the set of accepting states. The corresponding DFA has states corresponding to subsets of Q. The initial state of the DFA is {q0}, the (one-element) set of initial states. The transition function of the DFA maps a state S (representing a subset of Q) and an input symbol x to the set T(S,x) = ∪{T(q,x) | q ∈ S}, the set of all states that can be reached by an x-transition from a state in S. A state S of the DFA is an accepting state if and only if at least one member of S is an accepting state of the NFA. In the simplest version of the powerset construction, the set of all states of the DFA is the powerset of Q, the set of all possible subsets of Q. However, many states of the resulting DFA may be useless as they may be unreachable from the initial state. An alternative version of the construction creates only the states that are actually reachable. === NFA with ε-moves === For an NFA with ε-moves (also called an ε-NFA), the construction must be modified to deal with these by computing the ε-closure of states: the set of all states reachable from some given state using only ε-moves. Van Noord recognizes three possible ways of incorporating this closure computation in the powerset construction: Compute the ε-closure of the entire automaton as a preprocessing step, producing an equivalent NFA without ε-moves, then apply the regular powerset construction. This version, also discussed by Hopcroft and Ullman, is straightforward to implement, but impractical for automata with large numbers of ε-moves, as commonly arise in natural language processing application. During the powerset computation, compute the ε-closure { q ′ | q → ε ∗ q ′ } {\displaystyle \{q'~|~q\to _{\varepsilon }^{}q'\}} of each state q that is considered by the algorithm (and cache the result). During the powerset computation, compute the ε-closure { q ′ | ∃ q ∈ Q ′ , q → ε ∗ q ′ } {\displaystyle \{q'~|~\exists q\in Q',q\to _{\varepsilon }^{}q'\}} of each subset of states Q' that is considered by the algorithm, and add its elements to Q'. === Multiple initial states === If NFAs are defined to allow for multiple initial states, the initial state of the corresponding DFA is the set of all initial states of the NFA, or (if the NFA also has ε-moves) the set of all states reachable from initial states by ε-moves. == Example == The NFA below has four states; state 1 is initial, and states 3 and 4 are accepting. Its alphabet consists of the two symbols 0 and 1, and it has ε-moves. The initial state of the DFA constructed from this NFA is the set of all NFA states that are reachable from state 1 by ε-moves; that is, it is the set {1,2,3}. A transition from {1,2,3} by input symbol 0 must follow either the arrow from state 1 to state 2, or the arrow from state 3 to state 4. Additionally, neither state 2 nor state 4 have outgoing ε-moves. Therefore, T({1,2,3},0) = {2,4}, and by the same reasoning the full DFA constructed from the NFA is as shown below. As can be seen in this example, there are five states reachable from the start state of the DFA; the remaining 11 sets in the powerset of the set of NFA states are not reachable. == Complexity == Because the DFA states consist of sets of NFA states, an n-state NFA may be converted to a DFA with at most 2n states. For every n, there exist n-state NFAs such that every subset of states is reachable from the initial subset, so that the converted DFA has exactly 2n states, giving Θ(2n) worst-case time complexity. A simple example requiring nearly this many states is the language of strings over the alphabet {0,1} in which there are at least n characters, the nth from last of which is 1. It can be represented by an (n + 1)-state NFA, but it requires 2n DFA states, one for each n-character suffix of the input; cf. picture for n=4. == Applications == Brzozowski's algorithm for DFA minimization uses the powerset construction, twice. It converts the input DFA into an NFA for the reverse language, by reversing all its arrows and exchanging the roles of initial and accepting states, converts the NFA back into a DFA using the powerset construction, and then repeats its process. Its worst-case complexity is exponential, unlike some other known DFA minimization algorithms, but in many examples it performs more quickly than its worst-case complexity would suggest. Safra's construction, which converts a non-deterministic Büchi automaton with n states into a deterministic Muller automaton or into a deterministic Rabin automaton with 2O(n log n) states, uses the powerset construction as part of its machinery.
Read more →
AI Website Builders: Free vs Paid (2026)

Looking for the best AI website builder? An AI website builder is software that uses machine learning to help you get more done — it can save you hours every week by automating repetitive work. Most options offer a generous free tier, with paid plans unlocking higher limits, faster processing, and team features. Whether you are a beginner or a pro, the right AI website builder slots into your workflow and pays for itself fast. This guide breaks down the top picks, their pros and cons, and who each one is best for.
Read more →
Concept mining

Concept mining is an activity that results in the extraction of concepts from artifacts. Solutions to the task typically involve aspects of artificial intelligence and statistics, such as data mining and text mining. Because artifacts are typically a loosely structured sequence of words and other symbols (rather than concepts), the problem is nontrivial, but it can provide powerful insights into the meaning, provenance and similarity of documents. == Methods == Traditionally, the conversion of words to concepts has been performed using a thesaurus, and for computational techniques the tendency is to do the same. The thesauri used are either specially created for the task, or a pre-existing language model, usually related to Princeton's WordNet. The mappings of words to concepts are often ambiguous. Typically each word in a given language will relate to several possible concepts. Humans use context to disambiguate the various meanings of a given piece of text, where available machine translation systems cannot easily infer context. For the purposes of concept mining, however, these ambiguities tend to be less important than they are with machine translation, for in large documents the ambiguities tend to even out, much as is the case with text mining. There are many techniques for disambiguation that may be used. Examples are linguistic analysis of the text and the use of word and concept association frequency information that may be inferred from large text corpora. Recently, techniques that base on semantic similarity between the possible concepts and the context have appeared and gained interest in the scientific community. == Applications == === Detecting and indexing similar documents in large corpora === One of the spin-offs of calculating document statistics in the concept domain, rather than the word domain, is that concepts form natural tree structures based on hypernymy and meronymy. These structures can be used to generate simple tree membership statistics, that can be used to locate any document in a Euclidean concept space. If the size of a document is also considered as another dimension of this space then an extremely efficient indexing system can be created. This technique is currently in commercial use locating similar legal documents in a 2.5 million document corpus. === Clustering documents by topic === Standard numeric clustering techniques may be used in "concept space" as described above to locate and index documents by the inferred topic. These are numerically far more efficient than their text mining cousins, and tend to behave more intuitively, in that they map better to the similarity measures a human would generate.
Read more →
Optical braille recognition

Optical braille recognition is technology to capture and process images of braille characters into natural language characters. It is used to convert braille documents for people who cannot read them into text, and for preservation and reproduction of the documents. == History == In 1984, a group of researchers at the Delft University of Technology designed a braille reading tablet, in which a reading head with photosensitive cells was moved along set of rulers to capture braille text line-by-line. In 1988, a group of French researchers at the Lille University of Science and Technology developed an algorithm, called Lectobraille, which converted braille documents into plain text. The system photographed the braille text with a low-resolution CCD camera, and used spatial filtering techniques, median filtering, erosion, and dilation to extract the braille. The braille characters were then converted to natural language using adaptive recognition. The Lectobraille technique had an error rate of 1%, and took an average processing time of seven seconds per line. In 1993, a group of researchers from the Katholieke Universiteit Leuven developed a system to recognize braille that had been scanned with a commercially available scanner. The system, however, was unable to handle deformities in the braille grid, so well-formed braille documents were required. In 1999, a group at the Hong Kong Polytechnic University implemented an optical braille recognition technique using edge detection to translate braille into English or Chinese text. In 2001, Murray and Dais created a handheld recognition system, that scanned small sections of a document at once. Because of the small area scanned at once, grid deformation was less of an issue, and a simpler, more efficient algorithm was employed. In 2003, Morgavi and Morando designed a system to recognize braille characters using artificial neural networks. This system was noted for its ability to handle image degradation more successfully than other approaches. == Challenges == Many of the challenges to successfully processing braille text arise from the nature of braille documents. Braille is generally printed on solid-color paper, with no ink to produce contrast between the raised characters and the background paper. However, imperfections in the page can appear in a scan or image of the page. Many documents are printed inter-point, meaning they are double-sided. As such, the depressions of the braille of one side appear interlaid with the protruding braille of the other side. == Techniques == Some optical braille recognition techniques attempt to use oblique lighting and a camera to reveal the shadows of the depressions and protrusions of the braille. Others make use of commercially available document scanners.
Read more →
BFR algorithm

The BFR algorithm, named after its inventors Bradley, Fayyad and Reina, is a variant of k-means algorithm that is designed to cluster data in a high-dimensional Euclidean space. It makes a very strong assumption about the shape of clusters: they must be normally distributed about a centroid. The mean and standard deviation for a cluster may differ for different dimensions, but the dimensions must be independent. In other words, the data must take the shape of axis-aligned ellipses.
Read more →
Baum–Welch algorithm

In electrical engineering, statistical computing and bioinformatics, the Baum–Welch algorithm is a special case of the expectation–maximization algorithm used to find the unknown parameters of a hidden Markov model (HMM). It makes use of the forward-backward algorithm to compute the statistics for the expectation step. The Baum–Welch algorithm, the primary method for inference in hidden Markov models, is numerically unstable due to its recursive calculation of joint probabilities. As the number of variables grows, these joint probabilities become increasingly small, leading to the forward recursions rapidly approaching values below machine precision. == History == The Baum–Welch algorithm was named after its inventors Leonard E. Baum and Lloyd R. Welch. The algorithm and the Hidden Markov models were first described in a series of articles by Baum and his peers at the IDA Center for Communications Research, Princeton in the late 1960s and early 1970s. One of the first major applications of HMMs was to the field of speech processing. In the 1980s, HMMs were emerging as a useful tool in the analysis of biological systems and information, and in particular genetic information. They have since become an important tool in the probabilistic modeling of genomic sequences. == Description == A hidden Markov model describes the joint probability of a collection of "hidden" and observed discrete random variables. It relies on the assumption that the i-th hidden variable given the (i − 1)-th hidden variable is independent of previous hidden variables, and the current observation variables depend only on the current hidden state. The Baum–Welch algorithm uses the well known EM algorithm to find the maximum likelihood estimate of the parameters of a hidden Markov model given a set of observed feature vectors. Let X t {\displaystyle X_{t}} be a discrete hidden random variable with N {\displaystyle N} possible values (i.e. We assume there are N {\displaystyle N} states in total). We assume the P ( X t ∣ X t − 1 ) {\displaystyle P(X_{t}\mid X_{t-1})} is independent of time t {\displaystyle t} , which leads to the definition of the time-independent stochastic transition matrix A = { a i j } = P ( X t = j ∣ X t − 1 = i ) . {\displaystyle A=\{a_{ij}\}=P(X_{t}=j\mid X_{t-1}=i).} The initial state distribution (i.e. when t = 1 {\displaystyle t=1} ) is given by π i = P ( X 1 = i ) . {\displaystyle \pi _{i}=P(X_{1}=i).} The observation variables Y t {\displaystyle Y_{t}} can take one of K {\displaystyle K} possible values. We also assume the observation given the "hidden" state is time independent. The probability of a certain observation y i {\displaystyle y_{i}} at time t {\displaystyle t} for state X t = j {\displaystyle X_{t}=j} is given by b j ( y i ) = P ( Y t = y i ∣ X t = j ) . {\displaystyle b_{j}(y_{i})=P(Y_{t}=y_{i}\mid X_{t}=j).} Taking into account all the possible values of Y t {\displaystyle Y_{t}} and X t {\displaystyle X_{t}} , we obtain the N × K {\displaystyle N\times K} matrix B = { b j ( y i ) } {\displaystyle B=\{b_{j}(y_{i})\}} where b j {\displaystyle b_{j}} belongs to all the possible states and y i {\displaystyle y_{i}} belongs to all the observations. An observation sequence is given by Y = ( Y 1 = y 1 , Y 2 = y 2 , … , Y T = y T ) {\displaystyle Y=(Y_{1}=y_{1},Y_{2}=y_{2},\ldots ,Y_{T}=y_{T})} . Thus we can describe a hidden Markov chain by θ = ( A , B , π ) {\displaystyle \theta =(A,B,\pi )} . The Baum–Welch algorithm finds a local maximum for θ ∗ = a r g m a x θ ⁡ P ( Y ∣ θ ) {\displaystyle \theta ^{}=\operatorname {arg\,max} _{\theta }P(Y\mid \theta )} (i.e. the HMM parameters θ {\displaystyle \theta } that maximize the probability of the observation). === Algorithm === Set θ = ( A , B , π ) {\displaystyle \theta =(A,B,\pi )} with random initial conditions. They can also be set using prior information about the parameters if it is available; this can speed up the algorithm and also steer it toward the desired local maximum. ==== Forward procedure ==== Let α i ( t ) = P ( Y 1 = y 1 , … , Y t = y t , X t = i ∣ θ ) {\displaystyle \alpha _{i}(t)=P(Y_{1}=y_{1},\ldots ,Y_{t}=y_{t},X_{t}=i\mid \theta )} , the probability of seeing the observations y 1 , y 2 , … , y t {\displaystyle y_{1},y_{2},\ldots ,y_{t}} and being in state i {\displaystyle i} at time t {\displaystyle t} . This is found recursively: α i ( 1 ) = π i b i ( y 1 ) , {\displaystyle \alpha _{i}(1)=\pi _{i}b_{i}(y_{1}),} α i ( t + 1 ) = b i ( y t + 1 ) ∑ j = 1 N α j ( t ) a j i . {\displaystyle \alpha _{i}(t+1)=b_{i}(y_{t+1})\sum _{j=1}^{N}\alpha _{j}(t)a_{ji}.} Since this series converges exponentially to zero, the algorithm will numerically underflow for longer sequences. However, this can be avoided in a slightly modified algorithm by scaling α {\displaystyle \alpha } in the forward and β {\displaystyle \beta } in the backward procedure below. ==== Backward procedure ==== Let β i ( t ) = P ( Y t + 1 = y t + 1 , … , Y T = y T ∣ X t = i , θ ) {\displaystyle \beta _{i}(t)=P(Y_{t+1}=y_{t+1},\ldots ,Y_{T}=y_{T}\mid X_{t}=i,\theta )} that is the probability of the ending partial sequence y t + 1 , … , y T {\displaystyle y_{t+1},\ldots ,y_{T}} given starting state i {\displaystyle i} at time t {\displaystyle t} . We calculate β i ( t ) {\displaystyle \beta _{i}(t)} as, β i ( T ) = 1 , {\displaystyle \beta _{i}(T)=1,} β i ( t ) = ∑ j = 1 N β j ( t + 1 ) a i j b j ( y t + 1 ) . {\displaystyle \beta _{i}(t)=\sum _{j=1}^{N}\beta _{j}(t+1)a_{ij}b_{j}(y_{t+1}).} ==== Update ==== We can now calculate the temporary variables, according to Bayes' theorem: γ i ( t ) = P ( X t = i ∣ Y , θ ) = P ( X t = i , Y ∣ θ ) P ( Y ∣ θ ) = α i ( t ) β i ( t ) ∑ j = 1 N α j ( t ) β j ( t ) , {\displaystyle \gamma _{i}(t)=P(X_{t}=i\mid Y,\theta )={\frac {P(X_{t}=i,Y\mid \theta )}{P(Y\mid \theta )}}={\frac {\alpha _{i}(t)\beta _{i}(t)}{\sum _{j=1}^{N}\alpha _{j}(t)\beta _{j}(t)}},} which is the probability of being in state i {\displaystyle i} at time t {\displaystyle t} given the observed sequence Y {\displaystyle Y} and the parameters θ {\displaystyle \theta } ξ i j ( t ) = P ( X t = i , X t + 1 = j ∣ Y , θ ) = P ( X t = i , X t + 1 = j , Y ∣ θ ) P ( Y ∣ θ ) = α i ( t ) a i j β j ( t + 1 ) b j ( y t + 1 ) ∑ k = 1 N ∑ w = 1 N α k ( t ) a k w β w ( t + 1 ) b w ( y t + 1 ) , {\displaystyle \xi _{ij}(t)=P(X_{t}=i,X_{t+1}=j\mid Y,\theta )={\frac {P(X_{t}=i,X_{t+1}=j,Y\mid \theta )}{P(Y\mid \theta )}}={\frac {\alpha _{i}(t)a_{ij}\beta _{j}(t+1)b_{j}(y_{t+1})}{\sum _{k=1}^{N}\sum _{w=1}^{N}\alpha _{k}(t)a_{kw}\beta _{w}(t+1)b_{w}(y_{t+1})}},} which is the probability of being in state i {\displaystyle i} and j {\displaystyle j} at times t {\displaystyle t} and t + 1 {\displaystyle t+1} respectively given the observed sequence Y {\displaystyle Y} and parameters θ {\displaystyle \theta } . The denominators of γ i ( t ) {\displaystyle \gamma _{i}(t)} and ξ i j ( t ) {\displaystyle \xi _{ij}(t)} are the same ; they represent the probability of making the observation Y {\displaystyle Y} given the parameters θ {\displaystyle \theta } . The parameters of the hidden Markov model θ {\displaystyle \theta } can now be updated: π i ∗ = γ i ( 1 ) , {\displaystyle \pi _{i}^{}=\gamma _{i}(1),} which is the expected frequency spent in state i {\displaystyle i} at time 1 {\displaystyle 1} . a i j ∗ = ∑ t = 1 T − 1 ξ i j ( t ) ∑ t = 1 T − 1 γ i ( t ) , {\displaystyle a_{ij}^{}={\frac {\sum _{t=1}^{T-1}\xi _{ij}(t)}{\sum _{t=1}^{T-1}\gamma _{i}(t)}},} which is the expected number of transitions from state i to state j compared to the expected total number of transitions starting in state i, including from state i to itself. The number of transitions starting in state i is equivalent to the number of times state i is observed in the sequence from t = 1 to t = T − 1. b i ∗ ( v k ) = ∑ t = 1 T 1 y t = v k γ i ( t ) ∑ t = 1 T γ i ( t ) , {\displaystyle b_{i}^{}(v_{k})={\frac {\sum _{t=1}^{T}1_{y_{t}=v_{k}}\gamma _{i}(t)}{\sum _{t=1}^{T}\gamma _{i}(t)}},} where 1 y t = v k = { 1 if y t = v k , 0 otherwise {\displaystyle 1_{y_{t}=v_{k}}={\begin{cases}1&{\text{if }}y_{t}=v_{k},\\0&{\text{otherwise}}\end{cases}}} is an indicator function, and b i ∗ ( v k ) {\displaystyle b_{i}^{}(v_{k})} is the expected number of times the output observations have been equal to v k {\displaystyle v_{k}} while in state i {\displaystyle i} over the expected total number of times in state i {\displaystyle i} . These steps are now repeated iteratively until a desired level of convergence. Note: It is possible to over-fit a particular data set. That is, P ( Y ∣ θ final ) > P ( Y ∣ θ true ) {\displaystyle P(Y\mid \theta _{\text{final}})>P(Y\mid \theta _{\text{true}})} . The algorithm also does not guarantee a global maximum. ==== Multiple sequences ==== The algorithm described thus far assumes a single observed sequence Y = y 1 , … , y T {\displaystyle Y=y_{1},\ldots ,y_{T}} . However, in many situations, there are several sequences observed: Y 1 ,
Read more →
Display list

A display list, also called a command list in Direct3D 12 and a command buffer in Vulkan, is a series of graphics commands or instructions that are run when the list is executed. Systems that make use of display list functionality are called retained mode systems, while systems that do not are as opposed to immediate mode systems. In OpenGL, display lists are useful to redraw the same geometry or apply a set of state changes multiple times. This benefit is also used with Direct3D 12's bundle command lists. In Direct3D 12 and Vulkan, display lists are regularly used for per-frame recording and execution. == Origins in vector displays == The vector monitors or calligraphic displays of the 1960s and 1970s used electron beam deflection to draw line segments, points, and sometimes curves directly on a CRT screen. Because the image would immediately fade, it needed to be redrawn many times a second (storage tube CRTs retained the image until blanked, but they were unsuitable for interactive graphics). To refresh the display, a dedicated CPU called a Display Processor or Display Processing Unit (DPU) was used, which had a memory buffer for a "display list", "display file", or "display program" containing line segment coordinates and other information. Advanced Display Processors also supported control flow instructions, which were useful for drawing repetitive graphics such as text, and some could perform coordinate transformations such as 3D projection. == Home computer display list functionality == One of the earliest systems with a true display list was the Atari 8-bit computers. The display list (actually called so in Atari terminology) is a series of instructions for ANTIC, the video co-processor used in these machines. This program, stored in the computer's memory and executed by ANTIC in real-time, can specify blank lines, any of six text modes and eight graphics modes, which sections of the screen can be horizontally or vertically fine-scrolled, and trigger Display List Interrupts (called raster interrupts or HBI on other systems). The Amstrad PCW family contains a Display List function called the 'Roller RAM'. This is a 512-byte RAM area consisting of 256 16-bit pointers in RAM, one for each line of the 720 × 256 pixel display. Each pointer identifies the location of 90 bytes of monochrome pixels that hold the line's 720 pixel states. The 90 bytes of 8 pixel states are spaced at 8-byte intervals, so there are 7 unused bytes between each byte of pixel data. This suits how the text-orientated PCW constructs a typical screen buffer in RAM, where the first character's 8 rows are stored in the first 8 bytes, the second character's rows in the next 8 bytes, and so on. The Roller RAM was implemented to speed up display scrolling as it would have been unacceptably slow for its 3.4 MHz Z80 to move up the 23 KB display buffer 'by hand' i.e. in software. The Roller RAM starting entry used at the beginning of a screen refresh is controlled by a Z80-writable I/O register. Therefore, the screen can be scrolled simply by changing this I/O register. Another system using a Display List-like feature in hardware is the Amiga, which, not coincidentally, was also designed by some of the same people who developed the custom hardware for the Atari 8-bit computers. Once directed to produce a display mode, it would continue to do so automatically for every following scan line. The computer also included a dedicated co-processor, called "Copper", which ran a simple program or 'Copper List' intended for modifying hardware registers in sync with the display. The Copper List instructions could direct the Copper to wait for the display to reach a specific position on the screen, and then change the contents of hardware registers. In effect, it was a processor dedicated to servicing raster interrupts. The Copper was used by Workbench to mix multiple display modes (multiple resolutions and color palettes on the monitor at the same time), and by numerous programs to create rainbow and gradient effects on the screen. The Amiga Copper was also capable of reconfiguring the sprite engine mid-frame, with only one scanline of delay. This allowed the Amiga to draw more than its 8 hardware sprites, so long as the additional sprites did not share scanlines (or the one scanline gap) with more than 7 other sprites. i.e., so long as at least one sprite had finished drawing, another sprite could be added below it on the screen. Additionally, the later 32-bit AGA chipset allowed the drawing of bigger sprites (more pixels per row) while retaining the same multiplexing. The Amiga also had dedicated block-shifter ("blitter") hardware, which could draw larger objects into a framebuffer. This was often used in place of, or in addition to, sprites. In more primitive systems, the results of a display list can be simulated, though at the cost of CPU-intensive writes to certain display modes, color control, or other visual effect registers in the video device, rather than a series of rendering commands executed by the device. Thus, one must create the displayed image using some other rendering process, either before or while the CPU-driven display generation executes. In many cases, the image is also modified or re-rendered between frames. The image is then displayed in various ways, depending on the exact way in which the CPU-driven display code is implemented. Examples of the results possible on these older machines requiring CPU-driven video include effects such as Commodore 64/128's FLI mode, or Rainbow Processing on the ZX Spectrum. == Usage in OpenGL == To delimit a display list, the glNewList and glEndList functions are used, and to execute the list, the glCallList function is used. Almost all rendering commands that occur between the function calls are stored in the display list. Commands that affect the client state are not stored in display lists. Display lists are named with an integer value, and creating a display list with the same name as one already created overrides the first. The glNewList function expects two arguments: an integer representing the name of the list, and an enumeration for the compilation mode. The two modes include GL_COMPILE_AND_EXECUTE, which compiles and immediately executes, and GL_COMPILE, which only compiles the list. Display lists enable the use of the retained mode rendering pattern, which is a system in which graphics commands are recorded (retained) to execute in succession at a later time. This is contrary to immediate mode, where graphics commands are immediately executed on client calls. == Usage in Direct3D 12 == Command lists are created using the ID3D12Device::CreateCommandList function. Command lists may be created in several types: direct, bundle, compute, copy, video decode, video process, and video encoding. Direct command lists specify that a command list the GPU can execute, and doesn't inherit any GPU state. Bundles, are best used for storing and executing small sets of commands any number of times. This is used differently than regular command lists, where commands stored in a command list are typically executed only once. Compute command lists are used for general computations, with a common use being calculating mipmaps. A copy command list is strictly for copying and the video decode and video process command lists are for video decoding and processing respectively. Upon creation, command lists are in the recording state. Command lists may be re-used by calling the ID3D12GraphicsCommandList::Reset function. After recording commands, the command list must be transitioned out of the recording state by calling ID3D12GraphicsCommandList::Close. The command list is then executed by calling ID3D12CommandQueue::ExecuteCommandLists.
Read more →
Struc2vec

struc2vec is a framework to generate node vector representations on a graph that preserve the structural identity. In contrast to node2vec, that optimizes node embeddings so that nearby nodes in the graph have similar embedding, struc2vec captures the roles of nodes in a graph, even if structurally similar nodes are far apart in the graph. It learns low-dimensional representations for nodes in a graph, generating random walks through a constructed multi-layer graph starting at each graph node. It is useful for machine learning applications where the downstream application is more related with the structural equivalence of the nodes (e.g., it can be used to detect nodes in networks with similar functions, such as interns in the social network of a corporation). struc2vec identifies nodes that play a similar role based solely on the structure of the graph, for example computing the structural identity of individuals in social networks. In particular, struc2vec employs a degree-based method to measure the pairwise structural role similarity, which is then adopted to build the multi-layer graph. Moreover, the distance between the latent representation of nodes is strongly correlated to their structural similarity. The framework contains three optimizations: reducing the length of degree sequences considered, reducing the number of pairwise similarity calculations, and reducing the number of layers in the generated graph. struc2vec follows the intuition that random walks through a graph can be treated as sentences in a corpus. Each node in a graph is treated as an individual word, and short random walk is treated as a sentence. In its final phase, the algorithm employs Gensim's word2vec algorithm to learn embeddings based on biased random walks. Sequences of nodes are fed into a skip-gram or continuous bag of words model and traditional machine-learning techniques for classification can be used. It is considered a useful framework to learn node embeddings based on structural equivalence.
Read more →
How to Choose an AI Marketing Tool

Curious about the best AI marketing tool? An AI marketing tool is software that uses machine learning to help you get more done — it combines speed, accuracy, and an interface that just works. Hands-on testing shows real-world results vary, so a short free trial is the smartest way to decide. Whether you are a beginner or a pro, the right AI marketing tool slots into your workflow and pays for itself fast. This guide breaks down the top picks, their pros and cons, and who each one is best for.
Read more →