AI App Builder Free Unlimited

AI App Builder Free Unlimited — 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 →
Structured-light 3D scanner

A structured-light 3D scanner is a device used to capture the three-dimensional shape of an object by projecting light patterns, such as grids or stripes, onto its surface. The deformation of these patterns is recorded by cameras and processed using specialized algorithms to generate a detailed 3D model. Structured-light 3D scanning is widely employed in fields such as industrial design, quality control, cultural heritage preservation, augmented reality gaming, and medical imaging. Compared to laser-based 3D scanning, structured-light scanners use non-coherent light sources, such as LEDs or projectors, which enable faster data acquisition and eliminate potential safety concerns associated with lasers. However, the accuracy of structured-light scanning can be influenced by external factors, including ambient lighting conditions and the reflective properties of the scanned object. == Principle == Projecting a narrow band of light onto a three-dimensional surface creates a line of illumination that appears distorted when viewed from perspectives other than that of the projector. This distortion can be analyzed to reconstruct the geometry of the surface, a technique known as light sectioning. Projecting patterns composed of multiple stripes or arbitrary fringes simultaneously enables the acquisition of numerous data points at once, improving scanning speed. While various structured light projection techniques exist, parallel stripe patterns are among the most commonly used. By analyzing the displacement of these stripes, the three-dimensional coordinates of surface details can be accurately determined. === Generation of light patterns === Two major methods of stripe pattern generation have been established: Laser interference and projection. The laser interference method works with two wide planar laser beam fronts. Their interference results in regular, equidistant line patterns. Different pattern sizes can be obtained by changing the angle between these beams. The method allows for the exact and easy generation of very fine patterns with unlimited depth of field. Disadvantages are high cost of implementation, difficulties providing the ideal beam geometry, and laser typical effects like speckle noise and the possible self interference with beam parts reflected from objects. Typically, there is no means of modulating individual stripes, such as with Gray codes. The projection method uses incoherent light and basically works like a video projector. Patterns are usually generated by passing light through a digital spatial light modulator, typically based on one of the three currently most widespread digital projection technologies, transmissive liquid crystal, reflective liquid crystal on silicon (LCOS) or digital light processing (DLP; moving micro mirror) modulators, which have various comparative advantages and disadvantages for this application. Other methods of projection could be and have been used, however. Patterns generated by digital display projectors have small discontinuities due to the pixel boundaries in the displays. Sufficiently small boundaries however can practically be neglected as they are evened out by the slightest defocus. A typical measuring assembly consists of one projector and at least one camera. For many applications, two cameras on opposite sides of the projector have been established as useful. Invisible (or imperceptible) structured light uses structured light without interfering with other computer vision tasks for which the projected pattern will be confusing. Example methods include the use of infrared light or of extremely high framerates alternating between two exact opposite patterns. === Calibration === Geometric distortions by optics and perspective must be compensated by a calibration of the measuring equipment, using special calibration patterns and surfaces. A mathematical model is used for describing the imaging properties of projector and cameras. Essentially based on the simple geometric properties of a pinhole camera, the model also has to take into account the geometric distortions and optical aberration of projector and camera lenses. The parameters of the camera as well as its orientation in space can be determined by a series of calibration measurements, using photogrammetric bundle adjustment. === Analysis of stripe patterns === There are several depth cues contained in the observed stripe patterns. The displacement of any single stripe can directly be converted into 3D coordinates. For this purpose, the individual stripe has to be identified, which can for example be accomplished by tracing or counting stripes (pattern recognition method). Another common method projects alternating stripe patterns, resulting in binary Gray code sequences identifying the number of each individual stripe hitting the object. An important depth cue also results from the varying stripe widths along the object surface. Stripe width is a function of the steepness of a surface part, i.e. the first derivative of the elevation. Stripe frequency and phase deliver similar cues and can be analyzed by a Fourier transform. Finally, the wavelet transform has recently been discussed for the same purpose. In many practical implementations, series of measurements combining pattern recognition, Gray codes and Fourier transform are obtained for a complete and unambiguous reconstruction of shapes. Another method also belonging to the area of fringe projection has been demonstrated, utilizing the depth of field of the camera. It is also possible to use projected patterns primarily as a means of structure insertion into scenes, for an essentially photogrammetric acquisition. === Precision and range === The optical resolution of fringe projection methods depends on the width of the stripes used and their optical quality. It is also limited by the wavelength of light. An extreme reduction of stripe width proves inefficient due to limitations in depth of field, camera resolution and display resolution. Therefore, the phase shift method has been widely established: A number of at least 3, typically about 10 exposures are taken with slightly shifted stripes. The first theoretical deductions of this method relied on stripes with a sine wave shaped intensity modulation, but the methods work with "rectangular" modulated stripes, as delivered from LCD or DLP displays as well. By phase shifting, surface detail of e.g. 1/10 the stripe pitch can be resolved. Current optical stripe pattern profilometry hence allows for detail resolutions down to the wavelength of light, below 1 micrometer in practice or, with larger stripe patterns, to approx. 1/10 of the stripe width. Concerning level accuracy, interpolating over several pixels of the acquired camera image can yield a reliable height resolution and also accuracy, down to 1/50 pixel. Arbitrarily large objects can be measured with accordingly large stripe patterns and setups. Practical applications are documented involving objects several meters in size. Typical accuracy figures are: Planarity of a 2-foot (0.61 m) wide surface, to 10 micrometres (0.00039 in). Shape of a motor combustion chamber to 2 micrometres (7.9×10−5 in) (elevation), yielding a volume accuracy 10 times better than with volumetric dosing. Shape of an object 2 inches (51 mm) large, to about 1 micrometre (3.9×10−5 in) Radius of a blade edge of e.g. 10 micrometres (0.00039 in), to ±0.4 μm === Navigation === As the method can measure shapes from only one perspective at a time, complete 3D shapes have to be combined from different measurements in different angles. This can be accomplished by attaching marker points to the object and combining perspectives afterwards by matching these markers. The process can be automated, by mounting the object on a motorized turntable on robotic inspection cell, or CNC positioning device. Markers can as well be applied on a positioning device instead of the object itself. The 3D data gathered can be used to retrieve CAD (computer aided design) data and models from existing components (reverse engineering), hand formed samples or sculptures, natural objects or artifacts. === Challenges === As with all optical methods, reflective or transparent surfaces raise difficulties. Reflections cause light to be reflected either away from the camera or right into its optics. In both cases, the dynamic range of the camera can be exceeded. Transparent or semi-transparent surfaces also cause major difficulties. In these cases, coating the surfaces with a thin opaque lacquer just for measuring purposes is a common practice. A recent method handles highly reflective and specular objects by inserting a 1-dimensional diffuser between the light source (e.g., projector) and the object to be scanned. Alternative optical techniques have been proposed for handling perfectly transparent and specular objects. Double reflections and inter-reflections can cause the stripe pattern to be overlaid with unwanted ligh
Read more →
AdTruth

AdTruth is a software product and the digital media division of 41st Parameter, a company headquartered in Scottsdale, Arizona, with regional offices in San Jose, California; London, England; and Munich, Germany. AdTruth allows marketers to recognize and reach target audiences across online devices. AdTruth software identifies users for targeting, tracking, performance tracking across digital media, including mobile and desktop, by analysing patterns in large numbers of advertisements served over the internet, rather than through the use of cookies. == History == AdTruth was founded in 2011 by Ori Eisen of 41st Parameter, to repurpose the company's fraud detection and prevention technology, for use within the advertising industry to accurately target intended audiences, particularly in mobile. Eisen was joined by James Lamberti in the role of vice president and general manager. In 2012 41st Parameter raised $13 million in Series D financing from Norwest Venture Partners, Kleiner Perkins Caufield & Byers, Jafco Ventures and Georgian Partners, bringing total funding to about $35 million. In May 2012, AdTruth hosted a meeting of digital media executives to discuss Apple’s UDID deprecation, with the intent of developing a device-neutral replacement standard. AdTruth joined the World Wide Web Consortium's Tracking Protection Working Group, which provides guidance for implementing and adhering to Do Not Track policies. AdTruth also worked with privacy firm Truste to create a privacy compliant Do Not Track-style mechanism for mobile. In 2013, the company Experian purchased 41st Parameter, acquiring AdTruth as part of the deal. == Product == AdTruth software helps marketers track, target and retarget consumers using more than 100 parameters, including milliseconds in differences in the internal clock setting, to recognize a particular device anonymously. AdTruth's technology uses non-UDID information to identify a wide range of devices for cookieless ad targeting. Its technology currently has about a 90 percent accuracy rate on iOS, higher on Android and desktop. AdTruth also has mobile web to app bridging capabilities as well as DeviceInsight technology, enabling marketers to identify users across mobile web and app content. 41st Parameter's patented AdTruth technology is being used by MdotM, in response to the deprecation of the UDID that included tracking and targeting capabilities. == Competitors == AdTruth's main competitor is BlueCava, which deploys a similar device-fingerprinting technology.
Read more →
Huawei Member Center

Huawei Member Center is a benefits app which runs using Huawei Mobile Services. Originally launched in China, Huawei Member Center is now being developed primarily around devices such as P40 Pro and the Nova 7. == Membership Levels == The Huawei Member Center provides rewards in two primary ways, 1) device-specific & promotions and 2) via frequent use of Huawei products and apps, using points to redeem additional benefits. In China, Huawei members are already classified into three levels, the highest being “elite”. Membership level determines the level of perks received, from priority access to the service hotline, new device events & proprietary early-access opportunities. Huawei ran a number of member events in 2019 called "Huawei Member Day" to promote the Member Center including providing tips for the Mate 30 Pro and offering a 50Gb cloud storage upgrade to users. == HMC in China == Huawei Member Center Has seen significant adoption in China and the east, the rewards for use on the app have ranged from free book coupons, discounted travel and exclusive gifts of new devices, such as the Huawei Enjoy Z.
Read more →
Coghead

Coghead was a web application company based out of Redwood City, California. The company offered a web-based service for building and hosting custom online database applications. Applications were built around custom data collections and were typically designed to facilitate management of, and collaboration on, business data. Examples of Coghead's "gallery" applications include project management, simple Customer relationship management, bug tracking and extreme programming. Coghead's service was available through a limited-access beta program before "going live" for free trial accounts in April, 2007. Coghead launched a paid subscription plans in June, 2007. On February 19, 2009, Coghead announced that its intellectual property assets (its 'service') had been acquired by SAP AG (NYSE:SAP). == Product == Coghead's product was a fully hosted environment for building, accessing, and maintaining applications and the associated business data. Like other so-called "Web 2.0" companies, Coghead built its product around the idea of "software as a service". The product was intended to allow users to design a range of applications from scratch using only a drag and drop, WYSIWYG user interface, with very limited scripting or coding (if any) required. Coghead also offered its paid subscribers the ability to develop and publish "Coglets," web forms that allowed site visitors to view data in, or submit data into, the host's Coghead database. On February 19, 2009, Coghead announced that SAP AG had acquired the Coghead service through an asset purchase. The SAP asset purchase closed in the 1st Quarter 2009. Immediately upon closing the asset purchase, the public-facing service was taken off-line by SAP as they prepared to integrate the Coghead code with other SAP assets. This forced many of Coghead's customers to find alternative solutions.
Read more →
List of large language models

A large language model (LLM) is a type of machine learning model designed for natural language processing tasks such as language generation. LLMs are language models with many parameters, and are trained with self-supervised learning on a vast amount of text. == List == For the training cost column, 1 petaFLOP-day equals 1 petaFLOP/sec × 1 day, or 8.64×1019 FLOP (floating point operations). Only the cost of the largest model is shown. The number of parameters is measured in billions, and the training cost is measured in petaFLOP-days. === 2018 === === 2019 === === 2020 === === 2021 === === 2022 === === 2023 === === 2024 === === 2025 === === 2026 ===
Read more →
Contextual image classification

Contextual image classification, a topic of pattern recognition in computer vision, is an approach of classification based on contextual information in images. "Contextual" means this approach is focusing on the relationship of the nearby pixels, which is also called neighbourhood. The goal of this approach is to classify the images by using the contextual information. == Introduction == Similar as processing language, a single word may have multiple meanings unless the context is provided, and the patterns within the sentences are the only informative segments we care about. For images, the principle is same. Find out the patterns and associate proper meanings to them. As the image illustrated below, if only a small portion of the image is shown, it is very difficult to tell what the image is about. Even try another portion of the image, it is still difficult to classify the image. However, if we increase the contextual of the image, then it makes more sense to recognize. As the full images shows below, almost everyone can classify it easily. During the procedure of segmentation, the methods which do not use the contextual information are sensitive to noise and variations, thus the result of segmentation will contain a great deal of misclassified regions, and often these regions are small (e.g., one pixel). Compared to other techniques, this approach is robust to noise and substantial variations for it takes the continuity of the segments into account. Several methods of this approach will be described below. == Applications == === Functioning as a post-processing filter to a labelled image === This approach is very effective against small regions caused by noise. And these small regions are usually formed by few pixels or one pixel. The most probable label is assigned to these regions. However, there is a drawback of this method. The small regions also can be formed by correct regions rather than noise, and in this case the method is actually making the classification worse. This approach is widely used in remote sensing applications. === Improving the post-processing classification === This is a two-stage classification process: For each pixel, label the pixel and form a new feature vector for it. Use the new feature vector and combine the contextual information to assign the final label to the === Merging the pixels in earlier stages === Instead of using single pixels, the neighbour pixels can be merged into homogeneous regions benefiting from contextual information. And provide these regions to classifier. === Acquiring pixel feature from neighbourhood === The original spectral data can be enriched by adding the contextual information carried by the neighbour pixels, or even replaced in some occasions. This kind of pre-processing methods are widely used in textured image recognition. The typical approaches include mean values, variances, texture description, etc. === Combining spectral and spatial information === The classifier uses the grey level and pixel neighbourhood (contextual information) to assign labels to pixels. In such case the information is a combination of spectral and spatial information. === Powered by the Bayes minimum error classifier === Contextual classification of image data is based on the Bayes minimum error classifier (also known as a naive Bayes classifier). Present the pixel: A pixel is denoted as x 0 {\displaystyle x_{0}} . The neighbourhood of each pixel x 0 {\displaystyle x_{0}} is a vector and denoted as N ( x 0 ) {\displaystyle N(x_{0})} . The values in the neighbourhood vector is denoted as f ( x i ) {\displaystyle f(x_{i})} . Each pixel is presented by the vector ξ = ( f ( x 0 ) , f ( x 1 ) , … , f ( x k ) ) {\displaystyle \xi =\left(f(x_{0}),f(x_{1}),\ldots ,f(x_{k})\right)} x i ∈ N ( x 0 ) ; i = 1 , … , k {\displaystyle x_{i}\in N(x_{0});\quad i=1,\ldots ,k} The labels (classification) of pixels in the neighbourhood N ( x 0 ) {\displaystyle N(x_{0})} are presented as a vector η = ( θ 0 , θ 1 , … , θ k ) {\displaystyle \eta =\left(\theta _{0},\theta _{1},\ldots ,\theta _{k}\right)} θ i ∈ { ω 0 , ω 1 , … , ω k } {\displaystyle \theta _{i}\in \left\{\omega _{0},\omega _{1},\ldots ,\omega _{k}\right\}} ω s {\displaystyle \omega _{s}} here denotes the assigned class. A vector presents the labels in the neighbourhood N ( x 0 ) {\displaystyle N(x_{0})} without the pixel x 0 {\displaystyle x_{0}} η ^ = ( θ 1 , θ 2 , … , θ k ) {\displaystyle {\hat {\eta }}=\left(\theta _{1},\theta _{2},\ldots ,\theta _{k}\right)} The neighbourhood: Size of the neighbourhood. There is no limitation of the size, but it is considered to be relatively small for each pixel x 0 {\displaystyle x_{0}} . A reasonable size of neighbourhood would be 3 × 3 {\displaystyle 3\times 3} of 4-connectivity or 8-connectivity ( x 0 {\displaystyle x_{0}} is marked as red and placed in the centre). The calculation: Apply the minimum error classification on a pixel x 0 {\displaystyle x_{0}} , if the probability of a class ω r {\displaystyle \omega _{r}} being presenting the pixel x 0 {\displaystyle x_{0}} is the highest among all, then assign ω r {\displaystyle \omega _{r}} as its class. θ 0 = ω r if P ( ω r ∣ f ( x 0 ) ) = max s = 1 , 2 , … , R P ( ω s ∣ f ( x 0 ) ) {\displaystyle \theta _{0}=\omega _{r}\quad {\text{ if }}\quad P(\omega _{r}\mid f(x_{0}))=\max _{s=1,2,\ldots ,R}P(\omega _{s}\mid f(x_{0}))} The contextual classification rule is described as below, it uses the feature vector x 1 {\displaystyle x_{1}} rather than x 0 {\displaystyle x_{0}} . θ 0 = ω r if P ( ω r ∣ ξ ) = max s = 1 , 2 , … , R P ( ω s ∣ ξ ) {\displaystyle \theta _{0}=\omega _{r}\quad {\text{ if }}\quad P(\omega _{r}\mid \xi )=\max _{s=1,2,\ldots ,R}P(\omega _{s}\mid \xi )} Use the Bayes formula to calculate the posteriori probability P ( ω s ∣ ξ ) {\displaystyle P(\omega _{s}\mid \xi )} P ( ω s ∣ ξ ) = p ( ξ ∣ ω s ) P ( ω s ) p ( ξ ) {\displaystyle P(\omega _{s}\mid \xi )={\frac {p(\xi \mid \omega _{s})P(\omega _{s})}{p\left(\xi \right)}}} The number of vectors is the same as the number of pixels in the image. For the classifier uses a vector corresponding to each pixel x i {\displaystyle x_{i}} , and the vector is generated from the pixel's neighbourhood. The basic steps of contextual image classification: Calculate the feature vector ξ {\displaystyle \xi } for each pixel. Calculate the parameters of probability distribution p ( ξ ∣ ω s ) {\displaystyle p(\xi \mid \omega _{s})} and P ( ω s ) {\displaystyle P(\omega _{s})} Calculate the posterior probabilities P ( ω r ∣ ξ ) {\displaystyle P(\omega _{r}\mid \xi )} and all labels θ 0 {\displaystyle \theta _{0}} . Get the image classification result. == Algorithms == === Template matching === The template matching is a "brute force" implementation of this approach. The concept is first create a set of templates, and then look for small parts in the image match with a template. This method is computationally high and inefficient. It keeps an entire templates list during the whole process and the number of combinations is extremely high. For a m × n {\displaystyle m\times n} pixel image, there could be a maximum of 2 m × n {\displaystyle 2^{m\times n}} combinations, which leads to high computation. This method is a top down method and often called table look-up or dictionary look-up. === Lower-order Markov chain === The Markov chain also can be applied in pattern recognition. The pixels in an image can be recognised as a set of random variables, then use the lower order Markov chain to find the relationship among the pixels. The image is treated as a virtual line, and the method uses conditional probability. === Hilbert space-filling curves === The Hilbert curve runs in a unique pattern through the whole image, it traverses every pixel without visiting any of them twice and keeps a continuous curve. It is fast and efficient. === Markov meshes === The lower-order Markov chain and Hilbert space-filling curves mentioned above are treating the image as a line structure. The Markov meshes however will take the two dimensional information into account. === Dependency tree === The dependency tree is a method using tree dependency to approximate probability distributions.
Read more →
Stochastic parrot

In machine learning, the term stochastic parrot is a metaphor that frames large language models as systems that statistically mimic text without real understanding. The word "stochastic" – from the ancient Greek "στοχαστικός" (stokhastikos, 'based on guesswork') – is a term from probability theory meaning "randomly determined". The word "parrot" refers to parrots' ability to mimic human speech. The term was introduced in a 2021 paper on AI ethics titled "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜" and authored by Timnit Gebru, Emily M. Bender, Angelina McMillan-Major, and Margaret Mitchell. The paper outlined possible risks associated with large language models (LLMs). In December 2020, it was the subject of a workplace dispute between Gebru (then co-leader of Google's Ethical Artificial Intelligence Team) and Google, which had requested the retraction of the paper. The incident culminated in Gebru's controversial departure from the company. The paper was later presented at the 2021 ACM Conference, and the term "stochastic parrot" has seen widespread use in academic research concerning generative AI and LLMs. The term has been interpreted negatively as an insult towards AI. == Background == Timnit Gebru is an AI ethics researcher, Emily M. Bender is a linguist specializing in computational linguistics, and Margaret Mitchell is a computer scientist specializing in algorithmic bias. Gebru had joined Google in 2018, where she co-led a team on the ethics of artificial intelligence with Mitchell. In late 2020, the paper "On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜" was co-written by Gebru and five other researchers, four of whom were Google employees. The paper argues that large language models (LLMs) present significant risks such as environmental and financial costs, inscrutability leading to unknown dangerous biases, and potential for deception as LLMs do not understand the concepts underlying what they learn. The paper states that LLMs are "stitching together sequences of linguistic forms ... observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning." Therefore, they are labeled "stochastic parrots". === Dismissal of Gebru by Google === After the paper was submitted for consideration to the 2021 ACM Conference, Google requested that Gebru either retract the paper from the conference or remove the names of Google employees from it. Gebru refused to do so without further discussion, and emailed Google Research vice president Megan Kacholia that if the company could not explain the request for retraction and address other concerns regarding similar projects, she would plan to resign after a transition period, stating that they could "work on a last date". The following day, on December 2, 2020, Gebru received an email saying that Google was "accepting her resignation". Her abrupt firing sparked protests by Google employees and negative publicity for the company. == Usage == The phrase has been used by AI skeptics to signify that LLMs lack understanding of the meaning of their outputs. Sam Altman, CEO of OpenAI, used the term shortly after the release of ChatGPT in December 2022, tweeting "i am a stochastic parrot, and so r u". The term was nominated as the 2023 AI-related Word of the Year by the American Dialect Society. == Debate == Some LLMs, such as ChatGPT, have become capable of interacting with users in convincingly human-like conversations. The development of these new systems has deepened the discussion of the extent to which LLMs understand or are simply "parroting". According to machine learning researchers Lindholm, Wahlström, Lindsten, and Schön, the term "stochastic parrot" highlights two vital limitations of LLMs: LLMs are limited by the data they are trained on and are simply stochastically repeating contents of datasets. Because they are just making up outputs based on training data, LLMs do not understand if they are saying something incorrect or inappropriate. Lindholm et al. noted that, with poor quality datasets and other limitations, a learning machine might produce results that are "dangerously wrong". === Subjective experience === In the mind of a human being, words and language correspond to things one has experienced. For LLMs, according to proponents of the theory, words correspond only to other words and patterns of usage fed into their training data. Proponents of the idea of stochastic parrots thus conclude that statements about LLMs are due to "the human tendency to attribute meaning to text", and claim this occurs despite the LLMs not actually understanding language. === Fine-tuning === Kelsey Piper argued that the claim that LLMs are stochastic parrots or mere "next-token predictors" focuses on pre-training, ignoring that modern LLMs are also fine-tuned to follow instructions and to prefer accurate answers. === Hallucinations and mistakes === The tendency of LLMs to pass off false information as fact is held as support. Called hallucinations or confabulations, LLMs will occasionally synthesize information that matches some pattern. LLMs may fail to distinguish fact and fiction, which leads to the claim that they can't connect words to a comprehension of the world, as humans do. Furthermore, LLMs may fail to decipher complex or ambiguous grammar cases that rely on understanding the meaning of language. For example: The wet newspaper that fell down off the table is my favorite newspaper. But now that my favorite newspaper fired the editor I might not like reading it anymore. Can I replace 'my favorite newspaper' by 'the wet newspaper that fell down off the table' in the second sentence? GPT-4, an LLM released in March 2023, responded yes, not understanding that the meaning of "newspaper" is different in these two contexts; it is first an object and second an institution. === Benchmarks and experiments === One argument against the hypothesis that LLMs are stochastic parrot is their results on benchmarks for reasoning, common sense and language understanding. In 2023, some LLMs have shown good results on many language understanding tests, such as the Super General Language Understanding Evaluation (SuperGLUE). GPT-4 scored in the >90th-percentile on the Uniform Bar Examination and achieved 93% accuracy on the MATH benchmark of high-school Olympiad problems, results that exceed rote pattern-matching expectations. Such tests, and the smoothness of many LLM responses, help as many as 51% of AI professionals believe they can truly understand language with enough data, according to a 2022 survey. === Expert rebuttals === Some AI researchers dispute the notion that LLMs merely "parrot" their training data. Geoffrey Hinton, a pioneering figure in neural networks, counters that the metaphor misunderstands the prerequisite for accurate language prediction. He argues that "to predict the next word accurately, you have to understand the sentence", a view he presented on 60 Minutes in 2023. From this perspective, understanding is not an alternative to statistical prediction, but rather an emergent property required to perform it effectively at scale. Hinton also uses logical puzzles to demonstrate that LLMs actually understand language. A 2024 Scientific American investigation described a closed Berkeley workshop where state-of-the-art models solved novel tier-4 mathematics problems and produced coherent proofs, indicating reasoning abilities beyond memorization. The GPT-4 Technical Report showed human-level results on professional and academic exams (e.g., the Uniform Bar Exam and USMLE), challenging the "parrot" characterization. Anthropic conducted mechanistic interpretability research on Claude, using attribution graphs to identify circuits. The research showed how the LLM processes information via chains of fuzzy logical inference, and indicated an ability to plan ahead. They found that Claude 3.5 Haiku "employs remarkably general abstractions", forms "internally generated plans for its future outputs" and "works backwards from its longer-term goals". They noted that "The mechanisms of the model can apparently only be faithfully described using an overwhelmingly large causal graph." They also found that the model includes "mechanisms that could underlie a simple form of metacognition", in that it "thinks about" the level of its own knowledge before reaching its answer. === Interpretability === Another line of evidence against the 'stochastic parrot' claim comes from mechanistic interpretability, a research field dedicated to reverse-engineering LLMs to understand their internal workings. Rather than only observing the model's input-output behavior, these techniques probe the model's internal activations, which can be used to determine if they contain structured representations of the world. The goal is to investigate whether LLMs are merely manipulating surface statistics or if t
Read more →
Kdb+

kdb+ is a column-based relational time series database (TSDB) with in-memory (IMDB) abilities, developed and marketed by KX Systems. The database is commonly used in high-frequency trading (HFT) to store, analyze, process, and retrieve large data sets at high speed. kdb+ has the ability to handle billions of records and analyzes data within a database. The database is available in 32-bit and 64-bit versions for several operating systems. Financial institutions use kdb+ to analyze time series data such as stock or commodity exchange data. The database has also been used for other time-sensitive data applications including commodity markets such as energy trading, telecommunications, sensor data, log data, machine and computer network usage monitoring along with real time analytics in Formula One racing. == Overview == kdb+ is a high-performance column-store database that was designed to process and store large amounts of data. Commonly accessed data is pushed into random-access memory (RAM), which is faster to access than data in disk storage. Created with financial institutions in mind, the database was developed as a central repository to store time series data that supports real-time analysis of billions of records. kdb+ has the ability to analyze data over time and responds to queries similar to Structured Query Language (SQL). Columnar databases return answers to some queries in a more efficient way than row-based database management systems. kdb+ dictionaries, tables and nanosecond time stamps are native data types and are used to store time series data. At the core of kdb+ is the built-in programming language, q, a concise, expressive query array language, and dialect of the language APL. Q can manipulate streaming, real-time, and historical data. kdb+ uses q to aggregate and analyze data, perform statistical functions, and join data sets and supports SQL queries The vector language q was built for speed and expressiveness and eliminates most need for looping structures. kdb+ includes interfaces in C, C++, Java, C#, and Python. == History == In 1998, KX released kdb, a database built on the language K written by Arthur Whitney. In 2003, kdb+ was released as a 64-bit version of kdb. In 2004, the kdb+ tick market database framework was released along with kdb+ taq, a loader for the New York Stock Exchange (NYSE) taq data. kdb+ was created by Arthur Whitney, building on his prior work with array languages. In April 2007, KX announced that it was releasing a version of kdb+ for Mac OS X. Then, kdb+ was also available on the operating systems Linux, Windows, and Solaris. In September 2012, version 3.0 was released. It was optimized for Intel's upgraded processors with support for WebSockets, and universally unique identifiers (UUIDs, termed globally unique identifiers (GUID)s in Microsoft software). Intel's Advanced Vector Extensions (AVX) and Streaming SIMD Extensions 4 (SSE4) 4.2 on the Sandy Bridge processors of the time allowed for enhanced support of the kdb+ system. In June 2013, version 3.1 was released, with benchmarks up to 8 times faster than older versions. In March 2020, version 4.0 was released. New features included Multithreaded primitives, Intel Optane DC persistent memory support and Data at Rest Encryption.
Read more →
Mean shift

Mean shift is a non-parametric feature-space mathematical analysis technique for locating the maxima of a density function, a so-called mode-seeking algorithm. Application domains include cluster analysis in computer vision and image processing. == History == The mean shift procedure is usually credited to work by Fukunaga and Hostetler in 1975. It is, however, reminiscent of earlier work by Schnell in 1964. == Overview == Mean shift is a procedure for locating the maxima—the modes—of a density function given discrete data sampled from that function. This is an iterative method, and we start with an initial estimate x {\displaystyle x} . Let a kernel function K ( x i − x ) {\displaystyle K(x_{i}-x)} be given. This function determines the weight of nearby points for re-estimation of the mean. Typically a Gaussian kernel on the distance to the current estimate is used, K ( x i − x ) = e − c | | x i − x | | 2 {\displaystyle K(x_{i}-x)=e^{-c||x_{i}-x||^{2}}} . The weighted mean of the density in the window determined by K {\displaystyle K} is m ( x ) = ∑ x i ∈ N ( x ) K ( x i − x ) x i ∑ x i ∈ N ( x ) K ( x i − x ) {\displaystyle m(x)={\frac {\sum _{x_{i}\in N(x)}K(x_{i}-x)x_{i}}{\sum _{x_{i}\in N(x)}K(x_{i}-x)}}} where N ( x ) {\displaystyle N(x)} is the neighborhood of x {\displaystyle x} , a set of points for which K ( x i − x ) ≠ 0 {\displaystyle K(x_{i}-x)\neq 0} . The difference m ( x ) − x {\displaystyle m(x)-x} is called mean shift in Fukunaga and Hostetler. The mean-shift algorithm now sets x ← m ( x ) {\displaystyle x\leftarrow m(x)} , and repeats the estimation until m ( x ) {\displaystyle m(x)} converges. Although the mean shift algorithm has been widely used in many applications, a rigid proof for the convergence of the algorithm using a general kernel in a high dimensional space is still not known. Aliyari Ghassabeh showed the convergence of the mean shift algorithm in one dimension with a differentiable, convex, and strictly decreasing profile function. However, the one-dimensional case has limited real world applications. Also, the convergence of the algorithm in higher dimensions with a finite number of the stationary (or isolated) points has been proved. However, sufficient conditions for a general kernel function to have finite stationary (or isolated) points have not been provided. Gaussian Mean-Shift is an Expectation–maximization algorithm. == Details == Let data be a finite set S {\displaystyle S} embedded in the n {\displaystyle n} -dimensional Euclidean space, X {\displaystyle X} . Let K {\displaystyle K} be a flat kernel that is the characteristic function of the λ {\displaystyle \lambda } -ball in X {\displaystyle X} , In each iteration of the algorithm, s ← m ( s ) {\displaystyle s\leftarrow m(s)} is performed for all s ∈ S {\displaystyle s\in S} simultaneously. The first question, then, is how to estimate the density function given a sparse set of samples. One of the simplest approaches is to just smooth the data, e.g., by convolving it with a fixed kernel of width h {\displaystyle h} , where x i {\displaystyle x_{i}} are the input samples and k ( r ) {\displaystyle k(r)} is the kernel function (or Parzen window). h {\displaystyle h} is the only parameter in the algorithm and is called the bandwidth. This approach is known as kernel density estimation or the Parzen window technique. Once we have computed f ( x ) {\displaystyle f(x)} from the equation above, we can find its local maxima using gradient ascent or some other optimization technique. The problem with this "brute force" approach is that, for higher dimensions, it becomes computationally prohibitive to evaluate f ( x ) {\displaystyle f(x)} over the complete search space. Instead, mean shift uses a variant of what is known in the optimization literature as multiple restart gradient descent. Starting at some guess for a local maximum, y k {\displaystyle y_{k}} , which can be a random input data point x 1 {\displaystyle x_{1}} , mean shift computes the gradient of the density estimate f ( x ) {\displaystyle f(x)} at y k {\displaystyle y_{k}} and takes an uphill step in that direction. == Types of kernels == Kernel definition: Let X {\displaystyle X} be the n {\displaystyle n} -dimensional Euclidean space, R n {\displaystyle \mathbb {R} ^{n}} . The norm of x {\displaystyle x} is a non-negative number, ‖ x ‖ 2 = x ⊤ x ≥ 0 {\displaystyle \|x\|^{2}=x^{\top }x\geq 0} . A function K : X → R {\displaystyle K:X\rightarrow \mathbb {R} } is said to be a kernel if there exists a profile, k : [ 0 , ∞ ] → R {\displaystyle k:[0,\infty ]\rightarrow \mathbb {R} } , such that K ( x ) = k ( ‖ x ‖ 2 ) {\displaystyle K(x)=k(\|x\|^{2})} and k is non-negative. k is non-increasing: k ( a ) ≥ k ( b ) {\displaystyle k(a)\geq k(b)} if a < b {\displaystyle a Read more →
Pose (computer vision)

In the fields of computing and computer vision, pose (or spatial pose) represents the position and the orientation of an object, each usually in three dimensions. Poses are often stored internally as transformation matrices. The term “pose” is largely synonymous with the term “transform”, but a transform may often include scale, whereas pose does not. In computer vision, the pose of an object is often estimated from camera input by the process of pose estimation. This information can then be used, for example, to allow a robot to manipulate an object or to avoid moving into the object based on its perceived position and orientation in the environment. Other applications include skeletal action recognition. == Pose estimation == The specific task of determining the pose of an object in an image (or stereo images, image sequence) is referred to as pose estimation. Pose estimation problems can be solved in different ways depending on the image sensor configuration, and choice of methodology. Three classes of methodologies can be distinguished: Analytic or geometric methods: Given that the image sensor (camera) is calibrated and the mapping from 3D points in the scene and 2D points in the image is known. If also the geometry of the object is known, it means that the projected image of the object on the camera image is a well-known function of the object's pose. Once a set of control points on the object, typically corners or other feature points, has been identified, it is then possible to solve the pose transformation from a set of equations which relate the 3D coordinates of the points with their 2D image coordinates. Algorithms that determine the pose of a point cloud with respect to another point cloud are known as point set registration algorithms, if the correspondences between points are not already known. Genetic algorithm methods: If the pose of an object does not have to be computed in real-time a genetic algorithm may be used. This approach is robust especially when the images are not perfectly calibrated. In this particular case, the pose represent the genetic representation and the error between the projection of the object control points with the image is the fitness function. Learning-based methods: These methods use artificial learning-based system which learn the mapping from 2D image features to pose transformation. In short, this means that a sufficiently large set of images of the object, in different poses, must be presented to the system during a learning phase. Once the learning phase is completed, the system should be able to present an estimate of the object's pose given an image of the object. == Camera pose ==
Read more →
Convolutional neural network

A convolutional neural network (CNN) is a type of feedforward neural network that learns features via filter (or kernel) optimization. This type of deep learning network has been applied to process and make predictions from many different types of data including text, images and audio. CNNs are the de-facto standard in deep learning-based approaches to computer vision and image processing, and have only recently been replaced—in some cases—by newer architectures such as the transformer. Vanishing gradients and exploding gradients, seen during backpropagation in earlier neural networks, are prevented by the regularization that comes from using shared weights over fewer connections. For example, for each neuron in the fully-connected layer, 10,000 weights would be required for processing an image sized 100 × 100 pixels. However, applying cascaded convolution (or cross-correlation) kernels, only 25 weights for each convolutional layer are required to process 5x5-sized tiles. Higher-layer features are extracted from wider context windows, compared to lower-layer features. Some applications of CNNs include: image and video recognition, recommender systems, image classification, image segmentation, medical image analysis, natural language processing, brain–computer interfaces, and financial time series. CNNs are also known as shift invariant or space invariant artificial neural networks, based on the shared-weight architecture of the convolution kernels or filters that slide along input features and provide translation-equivariant responses known as feature maps. Counter-intuitively, most convolutional neural networks are not invariant to translation, due to the downsampling operation they apply to the input. Feedforward neural networks are usually fully connected networks, that is, each neuron in one layer is connected to all neurons in the next layer. The "full connectivity" of these networks makes them prone to overfitting data. Typical ways of regularization, or preventing overfitting, include: penalizing parameters during training (such as weight decay) or trimming connectivity (skipped connections, dropout, etc.) Robust datasets also increase the probability that CNNs will learn the generalized principles that characterize a given dataset rather than the biases of a poorly-populated set. Convolutional networks were inspired by biological processes in that the connectivity pattern between neurons resembles the organization of the animal visual cortex. Individual cortical neurons respond to stimuli only in a restricted region of the visual field known as the receptive field. The receptive fields of different neurons partially overlap such that they cover the entire visual field. CNNs use relatively little pre-processing compared to other image classification algorithms. This means that the network learns to optimize the filters (or kernels) through automated learning, whereas in traditional algorithms these filters are hand-engineered. This simplifies and automates the process, enhancing efficiency and scalability overcoming human-intervention bottlenecks. == Architecture == A convolutional neural network consists of an input layer, hidden layers and an output layer. In a convolutional neural network, the hidden layers include one or more layers that perform convolutions. Typically this includes a layer that performs a dot product of the convolution kernel with the layer's input matrix. This product is usually the Frobenius inner product, and its activation function is commonly ReLU. As the convolution kernel slides along the input matrix for the layer, the convolution operation generates a feature map, which in turn contributes to the input of the next layer. This is followed by other layers such as pooling layers, fully connected layers, and normalization layers. Here it should be noted how close a convolutional neural network is to a matched filter. === Convolutional layers === In a CNN, the input is a tensor with shape: (number of inputs) × (input height) × (input width) × (input channels) After passing through a convolutional layer, the image becomes abstracted to a feature map, also called an activation map, with shape: (number of inputs) × (feature map height) × (feature map width) × (feature map channels). Convolutional layers convolve the input and pass its result to the next layer. This is similar to the response of a neuron in the visual cortex to a specific stimulus. Each convolutional neuron processes data only for its receptive field. Although fully connected feedforward neural networks can be used to learn features and classify data, this architecture is generally impractical for larger inputs (e.g., high-resolution images), which would require massive numbers of neurons because each pixel is a relevant input feature. A fully connected layer for an image of size 100 × 100 has 10,000 weights for each neuron in the second layer. Convolution reduces the number of free parameters, allowing the network to be deeper. For example, using a 5 × 5 tiling region, each with the same shared weights, requires only 25 neurons. Using shared weights means there are many fewer parameters, which helps avoid the vanishing gradients and exploding gradients problems seen during backpropagation in earlier neural networks. To speed processing, standard convolutional layers can be replaced by depthwise separable convolutional layers, which are based on a depthwise convolution followed by a pointwise convolution. The depthwise convolution is a spatial convolution applied independently over each channel of the input tensor, while the pointwise convolution is a standard convolution restricted to the use of 1 × 1 {\displaystyle 1\times 1} kernels. === Pooling layers === Convolutional networks may include local and/or global pooling layers along with traditional convolutional layers. Pooling layers reduce the dimensions of data by combining the outputs of neuron clusters at one layer into a single neuron in the next layer. Local pooling combines small clusters, tiling sizes such as 2 × 2 are commonly used. Global pooling acts on all the neurons of the feature map. There are two common types of pooling in popular use: max and average. Max pooling uses the maximum value of each local cluster of neurons in the feature map, while average pooling takes the average value. === Fully connected layers === Fully connected layers connect every neuron in one layer to every neuron in another layer. It is the same as a traditional multilayer perceptron neural network (MLP). Each neuron in the fully connected layer receives input from all the neurons in the previous layer. These inputs are weighted and summed with the corresponding biases, and then passed through an activation function to perform a nonlinear transformation, generating the output. The flattened matrix goes through a fully connected layer to classify the images. === Receptive field === In neural networks, each neuron receives input from some number of locations in the previous layer. In a convolutional layer, each neuron receives input from only a restricted area of the previous layer called the neuron's receptive field. Typically the area is a square (e.g. 5 by 5 neurons). Whereas, in a fully connected layer, the receptive field is the entire previous layer. Thus, in each convolutional layer, each neuron takes input from a larger area in the input than previous layers. This is due to applying the convolution over and over, which takes the value of a pixel into account, as well as its surrounding pixels. When using dilated layers, the number of pixels in the receptive field remains constant, but the field is more sparsely populated as its dimensions grow when combining the effect of several layers. To manipulate the receptive field size as desired, there are some alternatives to the standard convolutional layer. For example, atrous or dilated convolution expands the receptive field size without increasing the number of parameters by interleaving visible and blind regions. Moreover, a single dilated convolutional layer can comprise filters with multiple dilation ratios, thus having a variable receptive field size. === Weights === Each neuron in a neural network computes an output value by applying a specific function to the input values received from the receptive field in the previous layer. The function that is applied to the input values is determined by a vector of weights and a bias (typically real numbers). Learning consists of iteratively adjusting these biases and weights. The vectors of weights and biases are called filters and represent particular features of the input (e.g., a particular shape). A distinguishing feature of CNNs is that many neurons can share the same filter. This reduces the memory footprint because a single bias and a single vector of weights are used across all receptive fields that share that filter, as opposed to each receptive field having its own bias and vector
Read more →
TinyML

TinyML (short for tiny machine learning) is an area of machine learning that focuses on deploying and running models on low-power, resource-constrained embedded systems such as microcontrollers and edge devices. TinyML supports on-device inference with low latency and minimal reliance on cloud connectivity, which makes it suitable for applications in the Internet of Things (IoT), wearable devices, and real-time systems. == History == The idea of running machine learning models on embedded systems has gained traction in the late 2010s, as model compression, quantization, and efficient neural network architectures progressed. The term TinyML was popularized in 2019 with the publication of the book TinyML by Pete Warden and Daniel Situnayake and the creation of the TinyML Foundation.
Read more →
Eigenmoments

EigenMoments is a set of orthogonal, noise robust, invariant to rotation, scaling and translation and distribution sensitive moments. Their application can be found in signal processing and computer vision as descriptors of the signal or image. The descriptors can later be used for classification purposes. It is obtained by performing orthogonalization, via eigen analysis on geometric moments. == Framework summary == EigenMoments are computed by performing eigen analysis on the moment space of an image by maximizing signal-to-noise ratio in the feature space in form of Rayleigh quotient. This approach has several benefits in Image processing applications: Dependency of moments in the moment space on the distribution of the images being transformed, ensures decorrelation of the final feature space after eigen analysis on the moment space. The ability of EigenMoments to take into account distribution of the image makes it more versatile and adaptable for different genres. Generated moment kernels are orthogonal and therefore analysis on the moment space becomes easier. Transformation with orthogonal moment kernels into moment space is analogous to projection of the image onto a number of orthogonal axes. Nosiy components can be removed. This makes EigenMoments robust for classification applications. Optimal information compaction can be obtained and therefore a few number of moments are needed to characterize the images. == Problem formulation == Assume that a signal vector s ∈ R n {\displaystyle s\in {\mathcal {R}}^{n}} is taken from a certain distribution having correlation C ∈ R n × n {\displaystyle C\in {\mathcal {R}}^{n\times n}} , i.e. C = E [ s s T ] {\displaystyle C=E[ss^{T}]} where E[.] denotes expected value. Dimension of signal space, n, is often too large to be useful for practical application such as pattern classification, we need to transform the signal space into a space with lower dimensionality. This is performed by a two-step linear transformation: q = W T X T s , {\displaystyle q=W^{T}X^{T}s,} where q = [ q 1 , . . . , q n ] T ∈ R k {\displaystyle q=[q_{1},...,q_{n}]^{T}\in {\mathcal {R}}^{k}} is the transformed signal, X = [ x 1 , . . . , x n ] T ∈ R n × m {\displaystyle X=[x_{1},...,x_{n}]^{T}\in {\mathcal {R}}^{n\times m}} a fixed transformation matrix which transforms the signal into the moment space, and W = [ w 1 , . . . , w n ] T ∈ R m × k {\displaystyle W=[w_{1},...,w_{n}]^{T}\in {\mathcal {R}}^{m\times k}} the transformation matrix which we are going to determine by maximizing the SNR of the feature space resided by q {\displaystyle q} . For the case of Geometric Moments, X would be the monomials. If m = k = n {\displaystyle m=k=n} , a full rank transformation would result, however usually we have m ≤ n {\displaystyle m\leq n} and k ≤ m {\displaystyle k\leq m} . This is specially the case when n {\displaystyle n} is of high dimensions. Finding W {\displaystyle W} that maximizes the SNR of the feature space: S N R t r a n s f o r m = w T X T C X w w T X T N X w , {\displaystyle SNR_{transform}={\frac {w^{T}X^{T}CXw}{w^{T}X^{T}NXw}},} where N is the correlation matrix of the noise signal. The problem can thus be formulated as w 1 , . . . , w k = a r g m a x w w T X T C X w w T X T N X w {\displaystyle {w_{1},...,w_{k}}=argmax_{w}{\frac {w^{T}X^{T}CXw}{w^{T}X^{T}NXw}}} subject to constraints: w i T X T N X w j = δ i j , {\displaystyle w_{i}^{T}X^{T}NXw_{j}=\delta _{ij},} where δ i j {\displaystyle \delta _{ij}} is the Kronecker delta. It can be observed that this maximization is Rayleigh quotient by letting A = X T C X {\displaystyle A=X^{T}CX} and B = X T N X {\displaystyle B=X^{T}NX} and therefore can be written as: w 1 , . . . , w k = a r g m a x x w T A w w T B w {\displaystyle {w_{1},...,w_{k}}={\underset {x}{\operatorname {arg\,max} }}{\frac {w^{T}Aw}{w^{T}Bw}}} , w i T B w j = δ i j {\displaystyle w_{i}^{T}Bw_{j}=\delta _{ij}} === Rayleigh quotient === Optimization of Rayleigh quotient has the form: max w R ( w ) = max w w T A w w T B w {\displaystyle \max _{w}R(w)=\max _{w}{\frac {w^{T}Aw}{w^{T}Bw}}} and A {\displaystyle A} and B {\displaystyle B} , both are symmetric and B {\displaystyle B} is positive definite and therefore invertible. Scaling w {\displaystyle w} does not change the value of the object function and hence and additional scalar constraint w T B w = 1 {\displaystyle w^{T}Bw=1} can be imposed on w {\displaystyle w} and no solution would be lost when the objective function is optimized. This constraint optimization problem can be solved using Lagrangian multiplier: max w w T A w {\displaystyle \max _{w}{w^{T}Aw}} subject to w T B w = 1 {\displaystyle {w^{T}Bw}=1} max w L ( w ) = max w ( w T A w − λ w T B w ) {\displaystyle \max _{w}{\mathcal {L}}(w)=\max _{w}(w{T}Aw-\lambda w^{T}Bw)} equating first derivative to zero and we will have: A w = λ B w {\displaystyle Aw=\lambda Bw} which is an instance of Generalized Eigenvalue Problem (GEP). The GEP has the form: A w = λ B w {\displaystyle Aw=\lambda Bw} for any pair ( w , λ ) {\displaystyle (w,\lambda )} that is a solution to above equation, w {\displaystyle w} is called a generalized eigenvector and λ {\displaystyle \lambda } is called a generalized eigenvalue. Finding w {\displaystyle w} and λ {\displaystyle \lambda } that satisfies this equations would produce the result which optimizes Rayleigh quotient. One way of maximizing Rayleigh quotient is through solving the Generalized Eigen Problem. Dimension reduction can be performed by simply choosing the first components w i {\displaystyle w_{i}} , i = 1 , . . . , k {\displaystyle i=1,...,k} , with the highest values for R ( w ) {\displaystyle R(w)} out of the m {\displaystyle m} components, and discard the rest. Interpretation of this transformation is rotating and scaling the moment space, transforming it into a feature space with maximized SNR and therefore, the first k {\displaystyle k} components are the components with highest k {\displaystyle k} SNR values. The other method to look at this solution is to use the concept of simultaneous diagonalization instead of Generalized Eigen Problem. === Simultaneous diagonalization === Let A = X T C X {\displaystyle A=X^{T}CX} and B = X T N X {\displaystyle B=X^{T}NX} as mentioned earlier. We can write W {\displaystyle W} as two separate transformation matrices: W = W 1 W 2 . {\displaystyle W=W_{1}W_{2}.} W 1 {\displaystyle W_{1}} can be found by first diagonalize B: P T B P = D B {\displaystyle P^{T}BP=D_{B}} . Where D B {\displaystyle D_{B}} is a diagonal matrix sorted in increasing order. Since B {\displaystyle B} is positive definite, thus D B > 0 {\displaystyle D_{B}>0} . We can discard those eigenvalues that large and retain those close to 0, since this means the energy of the noise is close to 0 in this space, at this stage it is also possible to discard those eigenvectors that have large eigenvalues. Let P ^ {\displaystyle {\hat {P}}} be the first k {\displaystyle k} columns of P {\displaystyle P} , now P T ^ B P ^ = D B ^ {\displaystyle {\hat {P^{T}}}B{\hat {P}}={\hat {D_{B}}}} where D B ^ {\displaystyle {\hat {D_{B}}}} is the k × k {\displaystyle k\times k} principal submatrix of D B {\displaystyle D_{B}} . Let W 1 = P ^ D B ^ − 1 / 2 {\displaystyle W_{1}={\hat {P}}{\hat {D_{B}}}^{-1/2}} and hence: W 1 T B W 1 = ( P ^ D B ^ − 1 / 2 ) T B ( P ^ D B ^ − 1 / 2 ) = I {\displaystyle W_{1}^{T}BW_{1}=({\hat {P}}{\hat {D_{B}}}^{-1/2})^{T}B({\hat {P}}{\hat {D_{B}}}^{-1/2})=I} . W 1 {\displaystyle W_{1}} whiten B {\displaystyle B} and reduces the dimensionality from m {\displaystyle m} to k {\displaystyle k} . The transformed space resided by q ′ = W 1 T X T s {\displaystyle q'=W_{1}^{T}X^{T}s} is called the noise space. Then, we diagonalize W 1 T A W 1 {\displaystyle W_{1}^{T}AW_{1}} : W 2 T W 1 T A W 1 W 2 = D A {\displaystyle W_{2}^{T}W_{1}^{T}AW_{1}W_{2}=D_{A}} , where W 2 T W 2 = I {\displaystyle W_{2}^{T}W_{2}=I} . D A {\displaystyle D_{A}} is the matrix with eigenvalues of W 1 T A W 1 {\displaystyle W_{1}^{T}AW_{1}} on its diagonal. We may retain all the eigenvalues and their corresponding eigenvectors since most of the noise are already discarded in previous step. Finally the transformation is given by: W = W 1 W 2 {\displaystyle W=W_{1}W_{2}} where W {\displaystyle W} diagonalizes both the numerator and denominator of the SNR, W T A W = D A {\displaystyle W^{T}AW=D_{A}} , W T B W = I {\displaystyle W^{T}BW=I} and the transformation of signal s {\displaystyle s} is defined as q = W T X T s = W 2 T W 1 T X T s {\displaystyle q=W^{T}X^{T}s=W_{2}^{T}W_{1}^{T}X^{T}s} . === Information loss === To find the information loss when we discard some of the eigenvalues and eigenvectors we can perform following analysis: η = 1 − t r a c e ( W 1 T A W 1 ) t r a c e ( D B − 1 / 2 P T A P D B − 1 / 2 ) = 1 − t r a c e ( D B ^ − 1 / 2 P ^ T A P ^ D B ^ − 1 / 2 ) t r a c e ( D B − 1 / 2 P T A P D B − 1 / 2 ) {\displaystyle {\begin{array}{lll}\eta &=&
Read more →
U-Net

U-Net is a convolutional neural network that was developed for image segmentation. The network is based on a fully convolutional neural network whose architecture was modified and extended to work with fewer training images and to yield more precise segmentation. Segmentation of a 512 × 512 image takes less than a second on a modern (2015) GPU using the U-Net architecture. The U-Net architecture has also been employed in diffusion models for iterative image denoising. This technology underlies many modern image generation models, such as DALL-E, Midjourney, and Stable Diffusion. U-Net is also being explored for language models. Tokenization is not a separate step, allowing the model to more easily understand spelling and concurrently vectorizing / tokenizing higher level concepts. == Description == The U-Net architecture stems from the so-called "fully convolutional network". The main idea is to supplement a usual contracting network by successive layers, where pooling operations are replaced by upsampling operators. Hence these layers increase the resolution of the output. A successive convolutional layer can then learn to assemble a precise output based on this information. One important modification in U-Net is that there are a large number of feature channels in the upsampling part, which allow the network to propagate context information to higher resolution layers. As a consequence, the expansive path is more or less symmetric to the contracting part, and yields a u-shaped architecture. The network only uses the valid part of each convolution without any fully connected layers. To predict the pixels in the border region of the image, the missing context is extrapolated by mirroring the input image. This tiling strategy is important to apply the network to large images, since otherwise the resolution would be limited by the GPU memory. Recently, there had also been an interest in receptive field based U-Net models for medical image segmentation. == Network architecture == The network consists of a contracting path and an expansive path, which gives it the u-shaped architecture. The contracting path is a typical convolutional network that consists of repeated application of convolutions, each followed by a rectified linear unit (ReLU) and a max pooling operation. During the contraction, the spatial information is reduced while feature information is increased. The expansive pathway combines the feature and spatial information through a sequence of up-convolutions and concatenations with high-resolution features from the contracting path. == Applications == There are many applications of U-Net in biomedical image segmentation, such as brain image segmentation (''BRATS'') and liver image segmentation ("siliver07") as well as protein binding site prediction. U-Net implementations have also found use in the physical sciences, for example in the analysis of micrographs of materials. Variations of the U-Net have also been applied for medical image reconstruction. Here are some variants and applications of U-Net as follows: Pixel-wise regression using U-Net and its application on pansharpening; 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation; TernausNet: U-Net with VGG11 Encoder Pre-Trained on ImageNet for Image Segmentation. Image-to-image translation to estimate fluorescent stains In binding site prediction of protein structure. == History == U-Net was created by Olaf Ronneberger, Philipp Fischer, Thomas Brox in 2015 and reported in the paper "U-Net: Convolutional Networks for Biomedical Image Segmentation". It is an improvement and development of FCN: Evan Shelhamer, Jonathan Long, Trevor Darrell (2014). "Fully convolutional networks for semantic segmentation".
Read more →