AI Face Or Not

AI Face Or Not — independent reviews, comparisons, pricing and step-by-step guides on Aizhi.

List of speech recognition software

Speech recognition software is available for many computing platforms, operating systems, use models, and software licenses. Here is a listing of such, grouped in various useful ways. == Acoustic models and speech corpus (compilation) == The following list presents notable speech recognition software engines with a brief synopsis of characteristics. == Macintosh == == Cross-platform web apps based on Chrome == The following list presents notable speech recognition software that operate in a Chrome browser as web apps. They make use of HTML5 Web-Speech-API. == Mobile devices and smartphones == Many mobile phone handsets, including feature phones and smartphones such as iPhones and BlackBerrys, have basic dial-by-voice features built in. Many third-party apps have implemented natural-language speech recognition support, including: == Windows == === Windows built-in speech recognition === The Windows Speech Recognition version 8.0 by Microsoft comes built into Windows Vista, Windows 7, Windows 8 and Windows 10. Speech Recognition is available only in English, French, Spanish, German, Japanese, Simplified Chinese, and Traditional Chinese and only in the corresponding version of Windows; meaning you cannot use the speech recognition engine in one language if you use a version of Windows in another language. Windows 7 Ultimate and Windows 8 Pro allow you to change the system language, and therefore change which speech engine is available. Windows Speech Recognition evolved into Cortana (software), a personal assistant included in Windows 10. === Windows 7, 8, 10, 11 third-party speech recognition === Braina – Dictate into third party software and websites, fill web forms and execute vocal commands. Dragon NaturallySpeaking from Nuance Communications – Successor to the older DragonDictate product. Focus on dictation. 64-bit Windows support since version 10.1. Tazti – Create speech command profiles to play PC games and control applications – programs. Create speech commands to open files, folders, webpages, applications. Windows 7, Windows 8 and Windows 8.1 versions. Voice Finger – software that improves the Windows speech recognition system by adding several extensions to it. The software enables controlling the mouse and the keyboard by only using the voice. It is especially useful for aiding users to overcome disabilities or to heal from computer injuries. === Microsoft Speech API === The first version of the Microsoft Speech API was released for Windows NT 3.51 and Windows 95 in 1995, it was then part of Windows up to Windows Vista. This initial version already contained Direct Speech Recognition and Direct Text To Speech APIs which applications could use to directly control engines, as well as simplified 'higher-level' Voice Command and Voice Talk APIs. Speech recognition functionality included as part of Microsoft Office and on Tablet PCs running Microsoft Windows XP Tablet PC Edition. It can also be downloaded as part of the Speech SDK 5.1 for Windows applications, but since that is aimed at developers building speech applications, the pure SDK form lacks any user interface (numerous applications were available), and thus is unsuitable for end users. == Built-in software == Microsoft Kinect includes built-in software which allows speech recognition of commands. Older generations of Nokia phones like Nokia N Series (before using Windows 7 mobile technology) used speech-recognition with family names from contact list and a few commands. Siri, originally implemented in the iPhone 4S, Apple's personal assistant for iOS, which uses technology from Nuance Communications. Cortana (software), Microsoft's personal assistant built into Windows Phone and Windows 10. == Interactive voice response == The following are interactive voice response (IVR) systems: CSLU Toolkit Genesys HTK – copyrighted by Microsoft, but allows altering software for licensee's internal use LumenVox ASR Tellme Networks; acquired by Microsoft == Unix-like x86 and x86-64 speech transcription software == Janus Recognition Toolkit (JRTk) Mozilla DeepSpeech was developing an open-source Speech-To-Text engine based on Baidu's deep speech research paper. Weesper Neon Flow – professional voice-dictation software that provides offline speech-to-text processing on macOS and Windows using local AI models. It is not open source and offers a paid subscription after a 15‑day free trial. Vocalinux – open-source speech transcription software for Linux. == Discontinued software == IBM VoiceType (formerly IBM Personal Dictation System) IBM ViaVoice – Embedded version still maintained by IBM. No longer supported for versions above Windows Vista. Untested above macOS 10.4 or on Macintoshes with an Intel chipset. Quack.com; acquired by AOL; the name has now been reused for an iPad search app. SpeechWorks from Nuance Communications. Yap Speech Cloud – Speech-to-text platform acquired by Amazon.com.
Read more →
Paola Velardi

Paola Velardi (born in Rome, April 26, 1955) is a full professor of computer science at Sapienza University in Rome, Italy. Her research encompasses Artificial Intelligence and specifically, natural language processing, machine learning business intelligence and semantic web. Velardi is one of the hundred female scientists included in the database "100esperte.it" (translated from Italian with "100 female experts"). This online, open database champions the recognition of top-rated female scientists in Science, Technology, Engineering and Mathematics (STEM) areas. Among her prestigious appointments and honors, her inclusion stands out —alongside 45 other international female scientists from the past, present, and future— in the Women in Science pavilion of UNESCO’s Virtual Science Museum. == Research == Paola Velardi's research activity has focused, since the early 1980s, on Artificial Intelligence, with a particular emphasis on natural language processing (NLP), Machine learning, and data mining. Her scientific contributions have evolved over time, following the sector's primary paradigms: Semantic Web and Ontologies: She is known for her pioneering work on semantic disambiguation and automated ontology learning, collaborating on the development of systems such as OntoLearn. Social Computing and Predictive Analysis: She has conducted research on extracting information from social media for epidemiological monitoring (syndromic surveillance) and for the identification of opinion leaders. In the educational field, she has developed machine learning models to predict the risk of student dropout. AI for Health and Elder Monitoring: She has coordinated projects to support frailty in the elderly, developing systems based on ambient intelligence and wearables to detect clinical and behavioral anomalies. She has also contributed to models for analyzing behavioral changes through dynamic clustering. Generative AI and Finance: More recently, her research has expanded into the use of generative AI and deep learning for finance, including benchmark studies on price trend prediction based on Limit Order Books (LOB) and the development of diffusion models for realistic market simulation (the TRADES project). According to Google Scholar bibliometrics updated until December 2025, Velardi's scientific publications have been cited more than 8100 times. Her h-index was 42. She has published more than 200 papers in international journals and conference proceedings. Some of her publications have been published in top rated journals such as Artificial Intelligence, Computational Linguistics, Knowledge-Based Systems, IEEE Transactions on Data and Knowledge Engineering , IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Transactions on Computers, IEEE Transactions on Software Engineering , Data Mining and Knowledge Discovery, and Journal of Web Semantics. == Education and previous employments == Velardi graduated in electronic engineering from Sapienza University in 1978. From 1978 to 1983, she worked for the Ugo Bordoni Foundation, a research institution focusing on ICT and working under the supervision of the Italian Ministry of Economic Development. In 1983, she was a visiting scholar at Stanford University. During this period she became passionate about Artificial Intelligence, which will remain her area of research throughout her career. From 1984 to 1986, she came back to her natal city and worked as a researcher for IBM. From 1986 to 1996 she was an associate professor in the engineering faculty of Polytechnic University of the Marches (Ancona, Italy). Starting in November 1996, she taught in and did research for the Department of Computer Science at the Sapienza University. Velardi was the head of Bachelor and Master Programs in Computer Science at Sapienza University from 2010 to 2013 and from 2015 to 2016. == Current employment == Since November 2001, Velardi has been a full professor in the department of computer science ("Dipartimento di Informatica" in Italian) at Sapienza University in Rome, Italy. Since 2013, she has been the coordinator of the Distance Learning Degree in Computer Science at Sapienza University. As of today, Velardi is a Senior Associate at the Institute of Cognitive Sciences and Technologies (ISTC) of the CNR. == Recognition == Velardi is one of the hundred female scientists included in the database "100esperte.it" (translated from Italian with "100 female experts"). This database lists top Italian female STEM scientists. Six out of one hundred scientists in the 100esperte's database are computer scientists like Velardi. Velardi is in the list of the top Italian scientists. A top scientist appearing in the Top-Italian-Scientists database is a scientist whose h-index is greater than 30. In March 2017, she was given an IBM Faculty Award for her research on social recommender systems. In December 2018, Velardi was included in the list of the 50 most influential Italian women in science and technology by Inspiring Fifty, a non-profit that aims to increase diversity in STEM by making female role models in tech more visible. In September 2019 she was the local co-organizer and Program Chair of the 6th ACM Celebration of Women in Computing. In November 2019 Velardi received the Standout Woman Award International at the seat of the Italian Parliament in Montecitorio. == Causes == Velardi aims at debunking the myth of computer science as a man-oriented and "inflexible" discipline. She is the founder of the project "NERD? Non e' roba per donne?" (translated from Italian: "NERD? Is it not stuff for women?"). This project was launched by Velardi in 2012 in the Department of Computer Science at Sapienza University. Since 2013 the project has been carried out in partnership with IBM Italy, which later created a spin-off of the project. The goal of the project is two-fold: (1) conveying computer science as creative, interdisciplinary and problem-solving-oriented science, and (2) encouraging young female students in studying computer science by, for instance, developing apps for smartphones. She has been the program chair of the 19th ACM celebration of Women in Computing. She is the creator and coordinator of the G4GRETA, an educational project that involves students of the third and fourth grades of Rome and Lazio. The project combines the development of IT skills with the themes of environmental sustainability and soft skills (teambuilding, pitching, social networking, etc.) Velardi is also involved in scientific dissemination. In 2020 and 2021 she cooperated with RaiCultura, the cultural division of RAI, the national broadcasting company.
Read more →
Top 10 Conversational AI Platforms Compared (2026)

In search of the best conversational AI platform? An conversational AI platform is software that uses machine learning to help you get more done — it turns a rough idea into a polished result in seconds. When choosing one, weigh output quality, pricing, export formats, and how well it fits the tools you already use. Whether you are a beginner or a pro, the right conversational AI platform 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 →
Best AI Bug Finders in 2026

In search of the best AI bug finder? An AI bug finder is software that uses machine learning to help you get more done — it turns a rough idea into a polished result in seconds. When choosing one, weigh output quality, pricing, export formats, and how well it fits the tools you already use. Whether you are a beginner or a pro, the right AI bug finder 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 →
Pixel shift

Pixel shift is a method in digital cameras for producing a super-resolution image. The method works by taking several images, after each such capture moving ("shifting") the sensor to a new position. In digital colour cameras that employ pixel shift, this avoids a major limitation inherent in using Bayer pattern for obtaining colour, and instead produces an image with increased colour resolution and, assuming a static subject or additional computational steps, an image free of colour moiré. Taking this idea further, sub-pixel shifting may increase the resolution of the final image beyond that suggested by the specified resolution of the image sensor. Additionally, assuming that the various individual captures are taken at the same sensitivity, the final combined image will have less image noise than a single capture. This can be thought of as an averaging effect (for instance, in a pixel shift image composed of four individual frames with a classic Bayer pattern, every pixel in the final colour image is based on two measurements of the green channel). == List of cameras implementing pixel shift == All of the following cameras are fabricated with one imaging sensor, thus any kind of pixel shift requires a movement of the whole sensor. === Canon === Canon R5: Contains a 45 Mpixel sensor. The High-Resolution Mode shifts the sensor by one pixel to obtain a sequence of nine images that are merged into a 400 Mpixel image. === Fujifilm === Fujifilm GFX50S II: contains a 51 Mpixel sensor. The Pixel Shift Multi-Shot mode shifts the imaging sensor by 0.5-pixel movements to obtain a sequence of 16 images that are subsequently merged into a 200 Mpixel image. Fujifilm GFX100, Fujifilm GFX100 II: contains a 102 Mpixel sensor. A sequence of 16 pixel shifted images are merged into a 400 Mpixel image. Fujifilm GFX100S, Fujifilm GFX100S II: contains a 102 Mpixel sensor. A sequence of 16 pixel shifted images are merged into a 400 Mpixel image Fujifilm GFX100IR: contains a 102 Mpixel sensor. A sequence of 16 pixel shifted images are merged into a 400 Mpixel image Fujifilm X-H2: contains a 40 Mpixel sensor. A sequence of 20 shifted images are merged into a 160 Mpixel image. Fujifilm X-T5: contains a 40 Mpixel sensor. A sequence of 20 shifted images are merged into a 160 Mpixel image. === Nikon === Nikon Z8: contains a 47.5 Mpixel sensor. The High Res shot mode shifts the imaging sensor by 0.5-pixel movements to obtain a sequence of up to 32 images that can be merged in Nikon's NX studio software. Nikon Zf: contains a 24 Mpixel sensor. The High Res shot mode shifts the imaging sensor by 0.5-pixel movements to obtain a sequence of up to 32 images that can be merged in Nikon's NX studio software. === Olympus === Olympus OM-D E-M1 Mark II: contains a 20.4 Mpixel sensor. The High Res shot mode produces a 50 Mpixel image. Olympus OM-D E-M5 Mark II: contains a 16 Mpixel sensor. The High Res shot mode shifts the imaging sensor by 0.5-pixel movements to obtain a sequence of 8 images that are subsequently merged into a 40 Mpixel image. Olympus OM-D E-M5 Mark III: contains a 20.4 Mpixel sensor. The High Res shot mode shifts the imaging sensor by 0.5-pixel movements to obtain a sequence of 8 images that are subsequently merged into a 50 Mpixel image. Olympus OM-D E-M1X: contains a 20.4 Mpixel sensor. The camera sports two pixel shift mode: (a) the 80Mp Tripod mode produces an 80 Mpixel image, (b) the Handheld High Res shot mode produces a 50 Mpixel image. Olympus PEN-F: contains a 20.4 Mpixel sensor. The High Res Shot mode takes multiple images, continually shifting the position of the sensor in sub-pixel increments. Combining these images results in either a 50MP JPEG or an 80MP Raw file. ==== OM System ==== OM System OM-1: contains a 20MPix sensor. The High Res Shot mode takes multiple images, and it can be used handheld or on a tripod. Handheld it will internally produce 50 Mpix files and 80 Mpix when mounted on a tripod. OM System OM-5: contains a 20MPix sensor. The High Res Shot mode takes multiple images, and it can be used handheld or on a tripod. Handheld it will internally produce 50 Mpix files and 80 Mpix when mounted on a tripod. === Panasonic === Panasonic Lumix DC-G9: contains a 20.3 Mpixel sensor. The High Resolution Mode takes a sequence of 8 shots in quick succession between which the sensor is shifted by 0.5 pixel for each image. These are subsequently merged into an 80 Mpixel image. Panasonic Lumix DC-S1: contains a 24.2 Mpixel sensor. The High Resolution Mode takes a sequence of shots in quick succession between which the sensor is shifted by a small amount. These are subsequently merged into a 96 Mpixel image. Panasonic Lumix DC-S1R: contains a 47.3 Mpixel sensor. The High Resolution Mode shifts the imaging sensor by a small increments to obtain a sequence of 8 images that are subsequently merged into a 187 Mpixel image. Panasonic Lumix DC-S1H Panasonic Lumix DC-S5 === Pentax === Pentax K-70: contains a 24.3 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into an image sporting 'all color data in each pixel to deliver super-high-resolution images'. Pentax KP: contains a 24.3 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into an image sporting 'high-resolution images with more accurate colours and much finer details'. Pentax K-3 II: contains a 24.3 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into an image sporting 'super-high-resolution images with far more truthful color reproduction and much finer details'. Pentax K-3 III: contains a 25.7 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into an image sporting 'a cancelling out of the Bayer pattern and removal of the need for sharpness-sapping demosaicing'. Pentax K-1: contains a 36.4 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into an image sporting 'improved detail and colour resolution'. Pentax K-1 II: contains a 36.4 Mpixel sensor. The camera sports two pixel shift mode: (a) a series of 4 tripod-stabilised images shifted by 1 pixel each are subsequently combined into a 47.3 Mpixel image, (b) a series of images taken in handheld mode are combined into a 47.3 Mpixel image that is, within limits, able to cope even with moving subjects. === Sony === Sony a6600: contains a 24.3 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into an image sporting 'all color data in each pixel to deliver super-high-resolution images'. Sony α7R III: contains a 42.4 Mpixel sensor. The pixel shift mode takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into a 42.4 Mpixel image with improved tonal resolution. Sony α7R IV: contains a 61 Mpixel sensor. The camera has two pixel shift modes, (a) the first takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into a 61 Mpixel image with improved tonal resolution, (b) the other takes a sequence of 16 shots between which the sensor is shifted by 0.5 pixel. These are subsequently merged into a 240 Mpixel image with both enhanced detail and improved tonal resolution. Sony α1: contains a 50 Mpixel sensor. The camera has two pixel shift modes, (a) the first takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into a 50 Mpixel image with improved tonal resolution, (b) the other takes a sequence of 16 shots between which the sensor is shifted by 0.5 pixel. These are subsequently merged into a 200 Mpixel image with both enhanced detail and improved tonal resolution. === Hasselblad === Hasselblad H3DII: the model H3DII-39 sports a 39 Mpixel sensor, the model H3DII-50 a 50 Mpixel sensor. Both enable a pixel shift mode which takes a sequence of 4 shots between which the sensor is shifted by 1 pixel. These are subsequently merged into a single image. Hasselblad H4D series: the model H4D-200MS contains a 50 Mpixel sensor. The sensor sports 3 different pixel shift modes which take (a) a sequence of 6 shots taken at slight offsets, (b) a sequence of 4 shots between which the sensor is shifted by 1 pixel, (c) a sequence of 4 shots between which the sensor is shifted by 0.5 pixels. Images obtained by all three modes are subsequently merged into 200 Mpixel images. Hasselblad H5D series: both models H5D-50c MS and H5D-200c MS contain a 50 Mpixel sensor. This sensor sports 2 different pixel shift modes which take (a) a sequence of 6 shots with full and half pixel moveme
Read more →
Top 10 AI Photo Editors Compared (2026)

Looking for the best AI photo editor? An AI photo editor 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 photo editor 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 →
Theano (software)

Theano is a Python library and optimizing compiler for manipulating and evaluating mathematical expressions, especially matrix-valued ones. In Theano, computations are expressed using a NumPy-esque syntax and compiled to run efficiently on either CPU or GPU architectures. == History == Theano is an open source project primarily developed by the Montreal Institute for Learning Algorithms (MILA) at the Université de Montréal. The name of the software references the ancient philosopher Theano, long associated with the development of the golden mean. On 28 September 2017, Pascal Lamblin posted a message from Yoshua Bengio, Head of MILA: major development would cease after the 1.0 release due to competing offerings by strong industrial players. Theano 1.0.0 was then released on 15 November 2017. On 17 May 2018, Chris Fonnesbeck wrote on behalf of the PyMC development team that the PyMC developers will officially assume control of Theano maintenance once the MILA development team steps down. On 29 January 2021, they started using the name Aesara for their fork of Theano. On 29 Nov 2022, the PyMC development team announced that the PyMC developers will fork the Aesara project under the name PyTensor. == Sample code == The following code is the original Theano's example. It defines a computational graph with 2 scalars a and b of type double and an operation between them (addition) and then creates a Python function f that does the actual computation. == Examples == === Matrix Multiplication (Dot Product) === The following code demonstrates how to perform matrix multiplication using Theano, which is essential for linear algebra operations in many machine learning tasks. === Gradient Calculation === The following code uses Theano to compute the gradient of a simple operation (like a neuron) with respect to its input. This is useful in training machine learning models (backpropagation). === Building a Simple Neural Network === The following code shows how to start building a simple neural network. This is a very basic neural network with one hidden layer. === Broadcasting in Theano === The following code demonstrates how broadcasting works in Theano. Broadcasting allows operations between arrays of different shapes without needing to explicitly reshape them.
Read more →
AI Code Generators: Free vs Paid (2026)

Looking for the best AI code generator? An AI code generator 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 code generator 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 →
Large language model

A large language model (LLM) is a neural network trained on a vast amount of text for natural language processing tasks, especially language generation. LLMs can typically generate, summarize, translate and analyze text in many contexts, and are a foundational technology behind modern chatbots. Biased or inaccurate training data can make an LLM's output less reliable. As of 2026, the most capable LLMs are based on transformer architectures, which, according to the 2017 paper "Attention Is All You Need", can be more efficient and parallelizable than earlier statistical and recurrent neural network models. Benchmark evaluations for LLMs attempt to measure model reasoning, factual accuracy, alignment, and safety. == History == Before the emergence of transformer-based models in 2017, some language models were considered large relative to the computational and data constraints of their time. In the early 1990s, IBM's statistical models pioneered word alignment techniques for machine translation, laying the groundwork for corpus-based language modeling. In 2001, a smoothed n-gram model, such as those employing Kneser–Ney smoothing, trained on 300 million words, achieved state-of-the-art perplexity on benchmark tests. During the 2000s, with the rise of widespread internet access, researchers began compiling massive text datasets from the web ("web as corpus") to train statistical language models. Moving beyond n-gram models, researchers started in 2000 to use neural networks as language models. Following the breakthrough of deep neural networks in image classification around 2012, similar architectures were adapted for language tasks. This shift was marked by the development of word embeddings (e.g., Word2Vec by Mikolov in 2013) and sequence-to-sequence (seq2seq) models using LSTM. In 2016, Google transitioned its translation service to neural machine translation (NMT), replacing statistical phrase-based models with deep recurrent neural networks. These early NMT systems used LSTM-based encoder-decoder architectures, as they preceded the invention of transformers. At the 2017 NeurIPS conference, Google researchers introduced the transformer architecture in their landmark paper "Attention Is All You Need". This paper's goal was to improve upon 2014 seq2seq technology, and was based mainly on the attention mechanism developed by Bahdanau et al. in 2014. The following year in 2018, BERT was introduced and quickly became "ubiquitous". Though the original transformer has both encoder and decoder blocks, BERT is an encoder-only model. Academic and research usage of BERT began to decline in 2023, following rapid improvements in the abilities of decoder-only models (such as GPT) to solve tasks via prompting. Although decoder-only GPT-1 was introduced in 2018, it was GPT-2 in 2019 that caught widespread attention because OpenAI claimed to have initially deemed it too powerful to release publicly, out of fear of malicious use. GPT-3 in 2020 went a step further and as of 2025 is available only via API with no offering of downloading the model to execute locally. But it was the consumer-facing chatbot ChatGPT in late 2022 that received extensive media coverage and public attention by 2023. The 2023 GPT-4 was praised for its increased accuracy and as a "holy grail" for its multimodal capabilities. OpenAI did not reveal the high-level architecture and the number of parameters of GPT-4. The release of ChatGPT led to an uptick in LLM usage across several research subfields of computer science, including robotics, software engineering, and societal impact work. In 2024, OpenAI released the reasoning model OpenAI o1, which generates long chains of thought before returning a final answer. Many LLMs with parameter counts comparable to those of OpenAI's GPT series have been developed. Since 2022, weights-available models have been gaining popularity, especially at first with BLOOM and LLaMA, though both have restrictions on usage and deployment. Mistral AI's open-weight models Mistral 7B and Mixtral 8x7B have a more permissive Apache License. In January 2025, DeepSeek released DeepSeek R1, a 671-billion-parameter open-weight model that performs comparably to OpenAI o1 but at a much lower price per token for users. Since 2023, many LLMs have been trained to be multimodal, having the ability to also process or generate other types of data, such as images, audio, or 3D meshes. Open-weight LLMs have become more influential since 2023. Per Vake et al. (2025), community-driven contributions to open-weight models improve their efficiency and performance via collaborative platforms such as Hugging Face. == Dataset preprocessing == === Tokenization === As machine learning algorithms process numbers rather than text, the text must be converted to numbers. In the first step, a vocabulary is decided upon, then integer indices are arbitrarily but uniquely assigned to each vocabulary entry, and finally, an embedding is associated with the integer index. Algorithms include byte-pair encoding (BPE) and WordPiece. There are also special tokens serving as control characters, such as [MASK] for masked-out token (as used in BERT), and [UNK] ("unknown") for characters not appearing in the vocabulary. Also, some special symbols are used to denote special text formatting. For example, "Ġ" denotes a preceding whitespace in RoBERTa and GPT and "##" denotes continuation of a preceding word in BERT. For example, the BPE tokenizer used by the legacy version of GPT-3 would split tokenizer: texts -> series of numerical "tokens" as Tokenization also compresses the datasets. Because LLMs generally require input to be an array that is not jagged, the shorter texts must be "padded" until they match the length of the longest one. ==== Byte-pair encoding ==== As an example, consider a tokenizer based on byte-pair encoding. In the first step, all unique characters (including blanks and punctuation marks) are treated as an initial set of n-grams (i.e. initial set of uni-grams). Successively the most frequent pair of adjacent characters is merged into a bi-gram and all instances of the pair are replaced by it. All occurrences of adjacent pairs of (previously merged) n-grams that most frequently occur together are then again merged into even lengthier n-gram, until a vocabulary of prescribed size is obtained. After a tokenizer is trained, any text can be tokenized by it, as long as it does not contain characters not appearing in the initial-set of uni-grams. === Dataset cleaning === In the context of training LLMs, datasets are typically cleaned by removing low-quality, duplicated, or toxic data. Cleaned datasets can increase training efficiency and lead to improved downstream performance. A trained LLM can be used to clean datasets for training a further LLM. With the increasing proportion of LLM-generated content on the web, data cleaning in the future may include filtering out such content. LLM-generated content can pose a problem if the content is similar to human text (making filtering difficult) but of lower quality (degrading performance of models trained on it). === Synthetic data === Training of largest language models might need more linguistic data than naturally available, or that the naturally occurring data is of insufficient quality. In these cases, synthetic data might be used. == Training == An LLM is a type of foundation model (large X model) trained on language. LLMs can be trained in different ways. In particular, GPT models are first pretrained to predict the next word on a large amount of data, before being fine-tuned. === Cost === Substantial infrastructure is necessary for training the largest models. The tendency towards larger models is visible in the list of large language models. For example, the training of GPT-2 (i.e. a 1.5-billion-parameter model) in 2019 cost $50,000, while training of the PaLM (i.e. a 540-billion-parameter model) in 2022 cost $8 million, and Megatron-Turing NLG 530B (in 2021) cost around $11 million. The qualifier "large" in "large language model" is inherently vague, as there is no definitive threshold for the number of parameters required to qualify as "large". === Fine-tuning === Before being fine-tuned, most LLMs are next-token predictors. The fine-tuning shapes the LLM's behavior via techniques like reinforcement learning from human feedback (RLHF) or constitutional AI. Instruction fine-tuning is a form of supervised learning used to teach LLMs to follow user instructions. In 2022, OpenAI demonstrated InstructGPT, a version of GPT-3 similarly fine-tuned to follow instructions. Reinforcement learning from human feedback (RLHF) involves training a reward model to predict which text humans prefer. Then, the LLM can be fine-tuned through reinforcement learning to better satisfy this reward model. Since humans typically prefer truthful, helpful and harmless answers, RLHF favors such answers. == Architecture == LLMs are generally based on the tra
Read more →
Michael Collins (computational linguist)

Michael J. Collins (born 4 March 1970) is a researcher in the field of computational linguistics. He is the Vikram S. Pandit Professor of Computer Science at Columbia University. His research interests are in natural language processing as well as machine learning and he has made important contributions in statistical parsing and in statistical machine learning. In his studies Collins covers a wide range of topics such as parse re-ranking, tree kernels, semi-supervised learning, machine translation and exponentiated gradient algorithms with a general focus on discriminative models and structured prediction. One notable contribution is a state-of-the-art parser for the Penn Wall Street Journal corpus. As of 11 November 2015, his works have been cited 16,020 times, and he has an h-index of 47. Collins worked as a researcher at AT&T Labs between January 1999 and November 2002, and later held the positions of assistant and associate professor at M.I.T. Since January 2011, he has been a professor at Columbia University. In 2011, he was named a fellow of the Association for Computational Linguistics.
Read more →
Isabelle Guyon

Isabelle Guyon (French pronunciation: [izabɛl ɡɥijɔ̃]; born August 15, 1961) is a French-born researcher in machine learning known for her work on support-vector machines, artificial neural networks and bioinformatics. She is a Chair Professor at the University of Paris-Saclay. Guyon serves as the Director of Research at Google DeepMind since October 2022. She is considered to be a pioneer in the field, with her contribution to the support-vector machines with Vladimir Vapnik and Bernhard Boser. == Biography == After graduating from the French engineering school ESPCI Paris in 1985, she joined the group of Gerard Dreyfus at the Université Pierre-et-Marie-Curie to do a PhD on neural networks architectures and training. Guyon defended her thesis in 1988 and was hired the year after at AT&T Bell Laboratories, first as a post-doc, then as a group leader. She worked at Bell Labs for six years, where she explored several research areas, from neural networks to pattern recognition and computational learning theory, with application to handwriting recognition. She collaborated with Yann LeCun, Léon Bottou, Vladimir Vapnik, Corinna Cortes, Yoshua Bengio, Patrice Simard, and met her future husband, Bernhard Boser. In 1996, Guyon left Bell Labs and raised her children at Berkeley, California. In Berkeley, she created her own machine learning consulting company, Clopinet. She became interested in medical applications, and used her previous work to classify the genes responsible for different types of cancers. Since 2003, Guyon has organized many challenges in data science, in order to stimulate research in this field. She founded ChaLearn in 2011, a non-profit organization aimed at creating machine learning challenges open to everyone. She was Program Chair of NeurIPS 2016 and became General Chair of NeurIPS in 2017. She is also Action Editor for the Journal of Machine Learning Research and Series Editor for Series: Challenges in Machine Learning. She is a member of the European Laboratory for Learning and Intelligent Systems. In 2016, Guyon came back to France to take the Chair Professorship in Big data between the University of Paris-Saclay and INRIA. She works in TAU (TAckling the Underspecified), a research collaboration of the Laboratoire de recherche en informatique. Together with Bernhard Schölkopf and Vladimir Vapnik, she received in 2020 the BBVA Foundation Frontiers of Knowledge Awards for her work in machine learning. == Scientific work == Guyon has worked in many subfields of machine learning, including neural networks, support-vector machines, feature selection and applications of machine learning to biology. === Support-vector machines === Among her most notable contributions, Guyon co-invented support-vector machines (SVM) in 1992, with Bernhard Boser and Vladimir Vapnik. SVM is a supervised machine learning algorithm, comparable to neural networks or decision trees, which has quickly become a classical technique in machine learning. SVMs have especially contributed to the popularization of kernel methods. === Neural networks === During her years at Bell Labs, Guyon took part of numerous projects involving neural networks. In particular, she wrote some of the first papers on the use of neural network for handwriting recognition using the MNIST database. She is also a co-inventor of the siamese neural networks, a neural network architecture used to learn similarities, with applications to signature, face or object recognition. === Machine learning for biology === Guyon is the author of many publications at the intersection of biology (cancer research and genomics) and artificial intelligence. She has notably introduced the use of support-vector machines to detect cancer using genes. === Machine learning challenges === Through her non-profit organization ChaLearn, Guyon has organized and directed challenges open to everyone in order to solve open problems in machine learning, including computer vision, neurosciences, particle physics, feature selection, causality and automated machine learning. Most of the challenges organized by ChaLearn have resulted in publications. Among the most cited ones are: Guyon et al., Result analysis of the NIPS 2003 feature selection challenge, Advances in neural information processing systems, 2005, link Escalera et al., ChaLearn Looking at People Challenge 2014: Dataset and Results, Computer Vision - ECCV 2014 Workshops, Springer International Publishing, 2014, link Guyon et al., A brief Review of the ChaLearn AutoML Challenge, JMLR: Workshop and Conference Proceedings 64:21-30, 2016, link Adam-Bourdario et al., The Higgs boson machine learning challenge, JMLR: Workshop and Conference Proceedings 42:19-55, 2015, link == Private life == She is married to Bernhard Boser, a professor at UC Berkeley. She has twins and one daughter, all three of whom have completed a science degree. Guyon has three citizenships: French by birth, Swiss by marriage and American by naturalization. == Awards and honors == Nomination at the French Academy of technologies (2024) Recipient of the BBVA Foundation Frontiers of Knowledge Awards (2020) American Medical Informatics Association Fellow (2011) == Publications == Bernhard Boser, Isabelle Guyon and Vladmir Vapnik, A training algorithm for optimal margin classifiers, Proceedings of the fifth annual workshop on Computational learning theory, 1992, doi:10.1145/130385.130401 Jane Bromley, Isabelle Guyon, Yann LeCun, Eduard Säckinger and Roopak Shah, Signature verification using a" siamese" time delay neural network, Advances in Neural Information Processing Systems, 1994. Isabelle Guyon and André Elisseeff, An introduction to variable and feature selection, Journal of Machine Learning Research, 2003. Isabelle Guyon, Jason Weston, Stephen Barnhill and Vladimir Vapnik, Gene selection for cancer classification using support vector machines, Machine Learning, Kluwer Academic Publishers, 2002, doi:10.1023/A:1012487302797
Read more →
Glushkov's construction algorithm

In computer science theory – particularly formal language theory – Glushkov's construction algorithm, invented by Victor Mikhailovich Glushkov, transforms a given regular expression into an equivalent nondeterministic finite automaton (NFA). Thus, it forms a bridge between regular expressions and nondeterministic finite automata: two abstract representations of the same class of formal languages. A regular expression may be used to conveniently describe an advanced search pattern in a "find and replace"–like operation of a text processing utility. Glushkov's algorithm can be used to transform it into an NFA, which furthermore is small by nature, as the number of its states equals the number of symbols of the regular expression, plus one. Subsequently, the NFA can be made deterministic by the powerset construction and then be minimized to get an optimal automaton corresponding to the given regular expression. The latter format is best suited for execution on a computer. From another, more theoretical point of view, Glushkov's algorithm is a part of the proof that NFA and regular expressions both accept exactly the same languages; that is, the regular languages. The converse of Glushkov's algorithm is Kleene's algorithm, which transforms a finite automaton into a regular expression. The automaton obtained by Glushkov's construction is the same as the one obtained by Thompson's construction algorithm, once its ε-transitions are removed. Glushkov's construction algorithm is also called The algorithm of Berry-Sethi, named after Gérard Berry and Ravi Sethi who worked on this construction. == Construction == Given a regular expression e, the Glushkov Construction Algorithm creates a non-deterministic automaton that accepts the language L ( e ) {\displaystyle L(e)} accepted by e. The construction uses four steps: === Step 1 === Linearisation of the expression. Each letter of the alphabet appearing in the expression e is renamed, so that each letter occurs at most once in the new expression e ′ {\displaystyle e'} . Glushkov's construction essentially relies on the fact that e ′ {\displaystyle e'} represents a local language L ( e ′ ) {\displaystyle L(e')} . Let A be the old alphabet and let B be the new one. === Step 2a === Computation of the sets P ( e ′ ) {\displaystyle P(e')} , D ( e ′ ) {\displaystyle D(e')} , and F ( e ′ ) {\displaystyle F(e')} . The first, P ( e ′ ) {\displaystyle P(e')} , is the set of letters which occurs as first letter of a word of L ( e ′ ) {\displaystyle L(e')} . The second, D ( e ′ ) {\displaystyle D(e')} , is the set of letters that can end a word of L ( e ′ ) {\displaystyle L(e')} . The last one, F ( e ′ ) {\displaystyle F(e')} , is the set of letter pairs that can occur in words of L ( e ′ ) {\displaystyle L(e')} , i.e. it is the set of factors of length two of the words of L ( e ′ ) {\displaystyle L(e')} . Those sets are mathematically defined by P ( e ′ ) = { x ∈ B ∣ x B ∗ ∩ L ( e ′ ) ≠ ∅ } {\displaystyle P(e')=\{x\in B\mid xB^{}\cap L(e')\neq \emptyset \}} , D ( e ′ ) = { y ∈ B ∣ B ∗ y ∩ L ( e ′ ) ≠ ∅ } {\displaystyle D(e')=\{y\in B\mid B^{}y\cap L(e')\neq \emptyset \}} , F ( e ′ ) = { u ∈ B 2 ∣ B ∗ u B ∗ ∩ L ( e ′ ) ≠ ∅ } {\displaystyle F(e')=\{u\in B^{2}\mid B^{}uB^{}\cap L(e')\neq \emptyset \}} . They are computed by induction over the structure of the expression, as explained below. === Step 2b === Computation of the set Λ ( e ′ ) {\displaystyle \Lambda (e')} which contains the empty word ε {\displaystyle \varepsilon } if this word belongs to L ( e ′ ) {\displaystyle L(e')} , and is the empty set otherwise. Formally, this is Λ ( e ′ ) = { ε } ∩ L ( e ′ ) {\displaystyle \Lambda (e')=\{\varepsilon \}\cap L(e')} . === Step 3 === Computation of automaton recognizing the local language, as defined by P ( e ′ ) {\displaystyle P(e')} , D ( e ′ ) {\displaystyle D(e')} , F ( e ′ ) {\displaystyle F(e')} , and Λ ( e ′ ) {\displaystyle \Lambda (e')} . By definition, the local language defined by the sets P, D, and F is the set of words which begin with a letter of P, end by a letter of D, and whose factors of length 2 belong to F, optionally also including the empty word; that is, it is the language: L ′ = ( P B ∗ ∩ B ∗ D ) ∖ B ∗ ( B 2 ∖ F ) B ∗ ∪ Λ ( e ′ ) {\displaystyle L'=(PB^{}\cap B^{}D)\setminus B^{}(B^{2}\setminus F)B^{}\cup \Lambda (e')} . Strictly speaking, it is the computation of the automaton for the local language denoted by this linearised expression that is Glushkov's construction. === Step 4 === Remove the linearisation, replacing each indexed letter B by the original letter of A. == Example == Consider the regular expression e = ( a ( a b ) ∗ ) ∗ + ( b a ) ∗ {\displaystyle e=(a(ab)^{})^{}+(ba)^{}} . == Computation of the set of letters == The computation of the sets P, D, F, and Λ is done inductively over the regular expression e ′ {\displaystyle e'} . One must give the values for ∅, ε (the symbols for the empty language and the singleton language containing the empty word), the letters, and the results of the operations + , ⋅ , ∗ {\displaystyle +,\cdot ,^{}} . The most costly operations are the cartesian products of sets for the computation of F. == Properties == The obtained automaton is non-deterministic, and it has as many states as the number of letters of the regular expression, plus one. It has been proven that every Thompson's automaton can be transformed into Glushkov's automaton via a ε-transitions elimination method. == Applications and deterministic expressions == The computation of the automaton by the expression occurs often; it has been systematically used in search functions, in particular by the Unix grep command. Similarly, XML's specification also uses such constructions; for more efficiency, regular expressions of a certain kind, called deterministic expressions, have been studied.
Read more →
Social software engineering

Social software engineering (SSE) is a branch of software engineering that is concerned with the social aspects of software development and the developed software. SSE focuses on the socialness of both software engineering and developed software. On the one hand, the consideration of social factors in software engineering activities, processes and CASE tools is deemed to be useful to improve the quality of both development process and produced software. Examples include the role of situational awareness and multi-cultural factors in collaborative software development. On the other hand, the dynamicity of the social contexts in which software could operate (e.g., in a cloud environment) calls for engineering social adaptability as a runtime iterative activity. Examples include approaches which enable software to gather users' quality feedback and use it to adapt autonomously or semi-autonomously. SSE studies and builds socially-oriented tools to support collaboration and knowledge sharing in software engineering. SSE also investigates the adaptability of software to the dynamic social contexts in which it could operate and the involvement of clients and end-users in shaping software adaptation decisions at runtime. Social context includes norms, culture, roles and responsibilities, stakeholder's goals and interdependencies, end-users perception of the quality and appropriateness of each software behaviour, etc. The participants of the 1st International Workshop on Social Software Engineering and Applications (SoSEA 2008) proposed the following characterization: Community-centered: Software is produced and consumed by and/or for a community rather than focusing on individuals Collaboration/collectiveness: Exploiting the collaborative and collective capacity of human beings Companionship/relationship: Making explicit the various associations among people Human/social activities: Software is designed consciously to support human activities and to address social problems Social inclusion: Software should enable social inclusion enforcing links and trust in communities Thus, SSE can be defined as "the application of processes, methods, and tools to enable community-driven creation, management, deployment, and use of software in online environments". One of the main observations in the field of SSE is that the concepts, principles, and technologies made for social software applications are applicable to software development itself as software engineering is inherently a social activity. SSE is not limited to specific activities of software development. Accordingly, tools have been proposed supporting different parts of SSE, for instance, social system design or social requirements engineering. Consequently vertical market software, such as software development tools, engineering tools, marketing tools or software that helps users in a decision-making process can profit from social components. Such vertical social software differentiates strongly in its user-base from traditional social software such as Yammer.
Read more →
Statistical machine translation

Statistical machine translation (SMT) is a machine translation approach where translations are generated on the basis of statistical models whose parameters are derived from the analysis of bilingual text corpora. The statistical approach contrasts with the rule-based approaches to machine translation as well as with example-based machine translation, that superseded the previous rule-based approach that required explicit description of each and every linguistic rule, which was costly, and which often did not generalize to other languages. The first ideas of statistical machine translation were introduced by Warren Weaver in 1949, including the ideas of applying Claude Shannon's information theory. Statistical machine translation was re-introduced in the late 1980s and early 1990s by researchers at IBM's Thomas J. Watson Research Center. Before the introduction of neural machine translation, it was by far the most widely studied machine translation method. == Basis == The idea behind statistical machine translation comes from information theory. A document is translated according to the probability distribution p ( e | f ) {\displaystyle p(e|f)} that a string e {\displaystyle e} in the target language (for example, English) is the translation of a string f {\displaystyle f} in the source language (for example, French). The problem of modeling the probability distribution p ( e | f ) {\displaystyle p(e|f)} has been approached in a number of ways. One approach which lends itself well to computer implementation is to apply Bayes' theorem, that is p ( e | f ) ∝ p ( f | e ) p ( e ) {\displaystyle p(e|f)\propto p(f|e)p(e)} , where the translation model p ( f | e ) {\displaystyle p(f|e)} is the probability that the source string is the translation of the target string, and the language model p ( e ) {\displaystyle p(e)} is the probability of seeing that target language string. This decomposition is attractive as it splits the problem into two subproblems. Finding the best translation e ~ {\displaystyle {\tilde {e}}} is done by picking up the one that gives the highest probability: e ~ = a r g max e ∈ e ∗ p ( e | f ) = a r g max e ∈ e ∗ p ( f | e ) p ( e ) {\displaystyle {\tilde {e}}=arg\max _{e\in e^{}}p(e|f)=arg\max _{e\in e^{}}p(f|e)p(e)} . For a rigorous implementation of this one would have to perform an exhaustive search by going through all strings e ∗ {\displaystyle e^{}} in the native language. Performing the search efficiently is the work of a machine translation decoder that uses the foreign string, heuristics and other methods to limit the search space and at the same time keeping acceptable quality. This trade-off between quality and time usage can also be found in speech recognition. As the translation systems are not able to store all native strings and their translations, a document is typically translated sentence by sentence. Language models are typically approximated by smoothed n-gram models, and similar approaches have been applied to translation models, but this introduces additional complexity due to different sentence lengths and word orders in the languages. Statistical translation models were initially word based (Models 1-5 from IBM Hidden Markov model from Stephan Vogel and Model 6 from Franz-Joseph Och), but significant advances were made with the introduction of phrase based models. Later work incorporated syntax or quasi-syntactic structures. == Benefits == The most frequently cited benefits of statistical machine translation (SMT) over rule-based approach are: More efficient use of human and data resources There are many parallel corpora in machine-readable format and even more monolingual data. Generally, SMT systems are not tailored to any specific pair of languages. More fluent translations owing to use of a language model == Shortcomings == Corpus creation can be costly. Specific errors are hard to predict and fix. Results may have superficial fluency that masks translation problems. Statistical machine translation usually works less well for language pairs with significantly different word order. The benefits obtained for translation between Western European languages are not representative of results for other language pairs, owing to smaller training corpora and greater grammatical differences. == Word-based translation == In word-based translation, the fundamental unit of translation is a word in some natural language. Typically, the number of words in translated sentences are different, because of compound words, morphology and idioms. The ratio of the lengths of sequences of translated words is called fertility, which tells how many foreign words each native word produces. Necessarily it is assumed by information theory that each covers the same concept. In practice this is not really true. For example, the English word corner can be translated in Spanish by either rincón or esquina, depending on whether it is to mean its internal or external angle. Simple word-based translation cannot translate between languages with different fertility. Word-based translation systems can relatively simply be made to cope with high fertility, such that they could map a single word to multiple words, but not the other way about. For example, if we were translating from English to French, each word in English could produce any number of French words— sometimes none at all. But there is no way to group two English words producing a single French word. An example of a word-based translation system is the freely available GIZA++ package (GPLed), which includes the training program for IBM models and HMM model and Model 6. The word-based translation is not widely used today; phrase-based systems are more common. Most phrase-based systems are still using GIZA++ to align the corpus. The alignments are used to extract phrases or deduce syntax rules. And matching words in bi-text is still a problem actively discussed in the community. Because of the predominance of GIZA++, there are now several distributed implementations of it online. == Phrase-based translation == In phrase-based translation, the aim is to reduce the restrictions of word-based translation by translating whole sequences of words, where the lengths may differ. The sequences of words are called blocks or phrases. These are typically not linguistic phrases, but phrasemes that were found using statistical methods from corpora. It has been shown that restricting the phrases to linguistic phrases (syntactically motivated groups of words, see syntactic categories) decreased the quality of translation. The chosen phrases are further mapped one-to-one based on a phrase translation table, and may be reordered. This table could be learnt based on word-alignment, or directly from a parallel corpus. The second model is trained using the expectation maximization algorithm, similarly to the word-based IBM model. == Syntax-based translation == Syntax-based translation is based on the idea of translating syntactic units, rather than single words or strings of words (as in phrase-based MT), i.e. (partial) parse trees of sentences/utterances. Until the 1990s, with advent of strong stochastic parsers, the statistical counterpart of the old idea of syntax-based translation did not take off. Examples of this approach include DOP-based MT and later synchronous context-free grammars. == Hierarchical phrase-based translation == Hierarchical phrase-based translation combines the phrase-based and syntax-based approaches to translation. It uses synchronous context-free grammar rules, but the grammars can be constructed by an extension of methods for phrase-based translation without reference to linguistically motivated syntactic constituents. This idea was first introduced in Chiang's Hiero system (2005). == Language models == A language model is an essential component of any statistical machine translation system, which aids in making the translation as fluent as possible. It is a function that takes a translated sentence and returns the probability of it being said by a native speaker. A good language model will for example assign a higher probability to the sentence "the house is small" than to "small the is house". Other than word order, language models may also help with word choice: if a foreign word has multiple possible translations, these functions may give better probabilities for certain translations in specific contexts in the target language. == Systems implementing statistical machine translation == Google Translate (started transition to neural machine translation in 2016) Microsoft Translator (started transition to neural machine translation in 2016) Yandex.Translate (switched to hybrid approach incorporating neural machine translation in 2017) == Challenges with statistical machine translation == Problems with statistical machine translation include: === Sentence alignment === Single sentences in one language can be found translated into several sentences in the o
Read more →
Writer invariant

Writer invariant, also called authorial invariant or author's invariant, is a property of a text which is invariant of its author, that is, it will be similar in all texts of a given author and different in texts of different authors. It can be used to find plagiarism or discover who is real author of anonymously published text. Writer invariant is also an author's pattern of writing a letter in handwritten text recognition. While it is generally recognised that writer invariants exist, it is not agreed what properties of a text should be used. Among the first ones used was distribution of word lengths; other proposed invariants include average sentence length, average word length, noun, verb or adjective usage frequency, vocabulary richness, and frequency of function words, or specific function words. Of these, average sentence lengths can be very similar in works of different authors or vary significantly even within a single work; average word lengths likewise turn out to be very similar in works of different authors. Analysis of function words shows promise because they are used by authors unconsciously.
Read more →