Preservation of documents, pictures, recordings, digital content, etc., is a major aspect of archival science. It is also an important consideration for people who are creating time capsules, family history, historical documents, scrapbooks and family trees. Common storage media are not permanent, and there are few reliable methods of preserving documents and pictures for the future. == Paper/prints (photos) == Color negatives and ordinary color prints may fade away to nothing in a relatively short period if not stored and handled properly. This happens even if the negatives and prints are kept in the dark, because ambient light is not the determining factor, but heat and humidity are. The color degradation is the result of the dyes used in the color processes. Because color processing results in a less stable image than traditional black-and-white processing, black-and-white pictures from the 1920s are more likely to survive long-term than color films and photographs from after the middle 20th century. Black-and-white photographic films using silver halide emulsions are the only film types that have proven to last for archival storage. The determining factors for longevity include the film base type, proper processing (develop, stop, fix and wash) and proper storage. Early films used a Cellulose nitrate base which was prone to decomposition and highly flammable. Nitrate film was replaced with acetate-base films. These Cellulose acetate films were later discovered to outgass acids (also referred to as vinegar syndrome). Acetate films were replaced in the early 1980s by polyester film base materials which have been determined to be more stable than film stocks with a nitrate or acetate base. Color prints made on most inkjet printers look very good at first but they have a very short lifespan, measured in months rather than in years. Even prints from commercial photo labs will start to fade in a matter of years if not processed properly and stored in cool, dry environments. == Documents/books == With documents for which the media are not so critical as what the documents contain, the information in documents can be copied by using photocopiers and image scanners. Books and manuscripts can also have their information saved without destruction by using a book scanner. Where the medium itself needs to be preserved, for example if a document is a crayon sketch by a famous artist on paper, a complex process of preservation may be used. Depending on the condition and importance of the item this can include gluing the media onto more stable media, or protective enclosing of the media. Polyester sleeves, acid-free folders, and pH buffered document boxes are common supportive protective enclosures whose selection must match the media's chemical and physical properties. Other considerations in preserving paper/books are: Damaging light, particularly UV light, which fades and destroys media over time by breaking down the molecules. Atmosphere contains small traces of sulfur dioxide and nitric acid which turn media yellow and break the fibers down. Humidity and moisture also aid in the breakdown of media. If there is too much, the document can be attacked by bacteria, and if too little, cellulose material breaks down. Temperature, particularly elevated ones, can destroy some media. Low temperatures can cause the water to form crystals which expands destroying the structure of paper-based documents. == Online photo albums == Although there are many websites that allow the upload of photographs and videos, digital preservation for the long-term is still considered an issue. There is a lack of confidence that such websites are capable of storing data for long periods of time (ex. 50 years) without data degradation or loss. == Optical media - CD, DVD, Blu-ray, M-Disc == Write-once optical media, such as CD-Rs and DVD-Rs, typically contain an organic dye that distinguishes data reading from data writing based on the dye's transparency along the disc. Conventional CDs and DVDs have finite shelf-life due to natural degradation of the dye; the newer M-DISC uses inorganic material technology to produce molded DVDs and Blu-Rays (up to 3-layer 100GB BDXL) with a claimed lifespan of 100-1000 years if stored correctly with most BD & BDXL rated read/writers enabling the higher power mode for the M-Disc format after 2011. The National Archives and Records Administration lists published life expectancies to be 10 or 25 years or more for normal CDs and DVDs and conservative life expectancies to be between 2 and 5 years. Storage environments, such as temperature and humidity, as well as handling conditions such as frequency of media use and compatibility between the recorder and media, affect media shelf-life. Improvements in media storage and migrations to new recording technologies can make certain formats obsolete within their respective lifespan. Technologists have pointed to internet streaming services, where services such as video-on-demand have contributed to the 33 percent decline in DVD sales the past 5 years, as a challenge for digital preservation. == Magnetic media - video cassettes, tapes, hard drives == Magnetic media such as audio and video tape and floppy disks also have limited life spans. Audio and video tapes require specific care and handling to ensure that the recorded information will be preserved. For information that must be preserved indefinitely, periodic transcription from old media to new ones is necessary, not only because the media are unstable but also because the recording technology may become obsolete. Magnetic media also deteriorates naturally with typical shelf lives between 10 and 20 years. Magnetic tape can degrade from binder hydrolysis or magnetic remanence decay. Binder hydrolysis, also known as sticky-shed syndrome, refers to the breakdown of binder, or glue, that holds the magnetic particles to the polyester base of the tape. Tapes which have been stored in hot, humid conditions are particularly vulnerable to this phenomenon and may suffer from accelerated degradation. Severe binder can cause the magnetic material to fall off or sheds from the base, leaving a pile of dust and clear backing. Archivists can bake the tape, which evaporates water molecules on the tape, to temporarily restore the binder before making a copy. Magnetic tape can also be destabilized by magnetic remanence decay, which refers to the weakening of the tape's magnetization over time. This weakens the affected tape's readability, leading to reduced sound clarity and volume or picture hue and contrast. Baking the tape will not restore magnetization. Media at risk include recorded media such as master audio recordings of symphonies and videotape recordings of the news gathered over the last 40 years. Threats to media that must be considered when archiving important record media include accidental erasure, physical loss due to disasters such as fires and floods, and media degradation. Along with the actual media being degraded over the years, the machines that are available to play back or reproduce the audio sources are becoming archaic themselves. Manufacturers and their support (parts, technical updates) for their machines have disappeared throughout the years. Even if the medium is vaulted and archived correctly, the mechanical properties of the machines have deteriorated to the point that they could do more harm than good to the tape being played. Many major film studios are now backing up their libraries by converting them to electronic media files, such as .AIFF or .WAV-based files via digital audio workstations. That way, even if the digital platform manufacturer goes out of business or no longer supports their product, the files can still be played on any common computer. There is a detailed process that must take place previous to the final archival product now that a digital solution is in place. Sample rates and their conversion and reference speed are both critical in this process. In floppy disks, the lubricants inside the plastic jackets of many older floppies promote the decay of the magnetic medium. Also, the alignment of the magnetic particles of the disk substrate may gradually degrade, leading to a loss of formatting and data. Early laser disk media were prone to degradation as the layers of the disk substrate were bonded with an adhesive that was vulnerable to decay and would crumble over time. This would lead the different layers of the disk to peel apart, damaging the pitted data surface and rendering the disk unreadable.
Google Messages
Google Messages (formerly known as Messenger, Android Messages, and Messages by Google) is a text messaging software application developed by Google for its Android and Wear OS mobile operating systems. It is also available as a web app. Google's official universal messaging platform for the Android ecosystem, Messages employs SMS, MMS, and Rich Communication Services (RCS). Starting in 2023, Google has RCS activated by default on participating Android devices, similar to the implementation of iMessage on Apple devices. Samsung Messages will be discontinued on July 6th 2026, with Samsung transitioning users to Google Messages as the default messaging application. == History == The original code for Android SMS messaging was released in 2009 integrated into the operating system. It was released as a standalone application independent of Android with the release of Android 5.0 Lollipop in 2014, replacing Google Hangouts as the default SMS app on Google's Nexus line of phones. In 2018, Messages adopted RCS messages and evolved to send larger data files, sync with other apps, and even create mass messages. This was in preparation for when Google launched Messages for web. In December 2019, Google began to introduce support for Rich Communication Services (RCS) messaging via an RCS service hosted by Google, referred to in the user interface as "chat features". This was followed by a wider global rollout throughout 2020. The app surpassed 1 billion installs in April 2020, doubling its number of installs in less than a year. Initially, RCS did not support end-to-end encryption. In June 2021, Google introduced end-to-end encryption in Messages by default using the Signal Protocol, for all one-to-one RCS-based conversations, for all RCS group chats in December 2022 for beta users, and for all RCS users by August 2023, as well as enabling RCS for all users by default to encourage encryption. In July 2023, Google announced it would build the Message Layer Security (MLS) end-to-end encryption protocol into Google Messages. Beginning with the Samsung Galaxy S21, Messages replaces Samsung's in-house Messages app as the default text messaging app for One UI for some regions and carriers. In April 2021, the app began to receive UI modifications on Samsung devices to follow aspects of One UI, including pushing the top of the message list towards the middle of the screen to improve ergonomics. In February 2023, Google began to replace references to "chat features" in the Messages user interface with "RCS". In August 2023, Google announced that Messages will use RCS by default for all users unless they opt out, to allow them to benefit from secure messaging. In December 2023, with the arrival of several new features, the app was renamed "Google Messages". In July 2024, Samsung announced it would no longer pre-install Samsung Messages on its Galaxy devices in some regions, starting with the Galaxy Z Fold 6 and Flip, favoring Google Messages instead. In April 2026, Samsung announced that Samsung Messages would be discontinued in July 2026. It encouraged users to switch to Google Messages. == Features == Some of the most important features in Google Messages are: Send instant text and voice messages in 1:1 or group chat conversations over mobile data and Wi-Fi, via Android, Wear OS or the web. End-to-end encryption for RCS chats. Typing, sent, delivered and read status Reply and react to specific messages Share files and high-resolution photos Voice message transcriptions Schedule messages In-app reminders for birthdays and messages you didn't respond to after some time with Nudges Tight integration with the Google ecosystem, e.g. Google Calendar, Meet, Maps, YouTube, Photos, Contacts, Assistant, Search, Safe Browsing etc. Web interface: Users can visit https://messages.google.com/web and either sign in with their Google account or scan the QR code that is shown with their smartphone to access a limited web version of the app that allows them to send and receive messages, provided the smartphone remains connected. Phone number recognition: The app shows the country and province of the caller. Additionally, it can show the company's name or a warning for spam calls if the number is registered in a data base. Access to the Gemini chatbot on select Pixel, Galaxy and Android devices.
Lexical choice
Lexical choice is the subtask of Natural language generation that involves choosing the content words (nouns, non-auxiliary verbs, adjectives, and adverbs) in a generated text. Function words (determiners, for example) are usually chosen during realisation. == Examples == The simplest type of lexical choice involves mapping a domain concept (perhaps represented in an ontology) to a word. For example, the concept Finger might be mapped to the word finger. A more complex situation is when a domain concept is expressed using different words in different situations. For example, the domain concept Value-Change can be expressed in many ways: The temperature rose: the verb rose is used for a Value-Change in temperature which increases the value. The temperature fell: the verb fell is used for a Value-Change in temperature which decreases the value. The rain got heavier: the phrase got heavier is used for a Value-Change in precipitation amount when the precipitation is rain. Sometimes words can communicate additional contextual information, for example: The temperature plummeted: the verb plummeted is used for a Value-Change in temperature which decreases the value, when the change is rapid and large. Contextual information is especially significant for vague terms such as tall. For example, a 2m tall man is tall, but a 2m tall horse is small. == Linguistic perspective == Lexical choice modules must be informed by linguistic knowledge of how the system's input data maps onto words. This is a question of semantics, but it is also influenced by syntactic factors (such as collocation effects) and pragmatic factors (such as context). Hence NLG systems need linguistic models of how meaning is mapped to words in the target domain (genre) of the NLG system. Genre tends to be very important; for example the verb veer has a very specific meaning in weather forecasts (wind direction is changing in a clockwise direction) which it does not have in general English, and a weather-forecast generator must be aware of this genre-specific meaning. In some cases there are major differences in how different people use the same word; for example, some people use by evening to mean 6PM and others use it to mean midnight. Psycholinguists have shown that when people speak to each other, they agree on a common interpretation via lexical alignment; this is not something which NLG systems can yet do. Ultimately, lexical choice must deal with the fundamental issue of how language relates to the non-linguistic world. For example, a system which chose colour terms such as red to describe objects in a digital image would need to know which RGB pixel values could generally be described as red; how this was influenced by visual (lighting, other objects in the scene) and linguistic (other objects being discussed) context; what pragmatic connotations were associated with red (for example, when an apple is called red, it is assumed to be ripe as well as have the colour red); and so forth. == Algorithms and models == A number of algorithms and models have been developed for lexical choice in the research community, for example Edmonds developed a model for choosing between near-synonyms (words with similar core meanings but different connotations). However such algorithms and models have not been widely used in applied NLG systems; such systems have instead often used quite simple computational models, and invested development effort in linguistic analysis instead of algorithm development.
Netomi
Netomi, formerly msg.ai, is an American artificial intelligence company and developer of chatbot technologies. == History == msg.ai was founded in May 2015 by Puneet Mehta. msg.ai worked with Sony Pictures to launch a chat bot on Facebook Messenger for a $100M film, Goosebumps and subsequently joined Y Combinator as a member of the Winter 2016 class. Later that year and in 2017, msg.ai completed two rounds of seed funding, led by Y Combinator and Index Ventures. In 2018, the company changed its name to Netomi. In 2019, the company raised $14.7 million in a Series A funding round also led by Index Ventures. In 2021, the company raised $30 million in a Series B funding round led by WndrCo LLC.
Elastix (image registration)
Elastix is an image registration toolbox built upon the Insight Segmentation and Registration Toolkit (ITK). It is entirely open-source and provides a wide range of algorithms employed in image registration problems. Its components are designed to be modular to ease a fast and reliable creation of various registration pipelines tailored for case-specific applications. It was first developed by Stefan Klein and Marius Staring under the supervision of Josien P.W. Pluim at Image Sciences Institute (ISI). Its first version was command-line based, allowing the final user to employ scripts to automatically process big data-sets and deploy multiple registration pipelines with few lines of code. Nowadays, to further widen its audience, a version called SimpleElastix is also available, developed by Kasper Marstal, which allows the integration of elastix with high level languages, such as Python, Java, and R. == Image registration fundamentals == Image registration is a well-known technique in digital image processing that searches for the geometric transformation that, applied to a moving image, obtains a one-to-one map with a target image. Generally, the images acquired from different sensors (multimodal), time instants (multitemporal), and points of view (multiview) should be correctly aligned to proceed with further processing and feature extraction. Even though there are a plethora of different approaches to image registration, the majority is composed of the same macro building blocks, namely the transformation, the interpolator, the metric, and the optimizer. Registering two or more images can be framed as an optimization problem that requires multiple iterations to converge to the best solution. Starting from an initial transformation computed from the image moments the optimization process searches for the best transformation parameters based on the value of the selected similarity metric. The figure on the right shows the high-level representation of the registration of two images, where the reference remains constant during the entire process, while the moving one will be transformed according to the transformation parameters. In other words, the registration ends when the similarity metric, which is a mathematical function with a certain number of parameters to be optimized, reaches the optimal value which is highly dependent on the specific application. == Main building blocks == Following the structure of the image registration workflow, the elastix toolbox proposes a modular solution that implements for each of the building blocks different algorithms, highly employed in medical image registration, and helps the final users to build their specific pipeline by selecting the most suitable algorithm for each of the main building blocks. Each block is easily configurable both by selecting pre-defined initialization values or by trying multiple sets of parameters and then choosing the most performing one. The registration is performed on images, and the elastix toolbox supports all the data formats supported by ITK, ranging from JPEG and PNG to medical standard formats such as DICOM and NIFTI. It also stores physical pixel spacing, the origin and the relative position to an external world reference system, when provided in the metadata, to facilitate the registration process, especially in medical field applications. === Transformation === The transformation is an essential building block, since it defines the allowable transformations. In image registration, the main distinction can be done between parallel-to-parallel and parallel-to-non parallel (deformable) line mapping transformations. In the elastix toolbox, the final users can select one transformation or compose more transformations either through addition or via composition. Below are reported the different transformation models in order of increasing flexibility, along with the corresponding elastix class names between brackets. Translation (TranslationTransform) allows only translations Rigid (EulerTransform) expands the translation adding rotations and the object is seen as a rigid body Similarity (SimilarityTransform) expands the rigid transformation by introducing isotropic scaling Affine (AffineTransform) expands the rigid transformation allowing both scaling and shear B-splines (BSplineTransform) is a deformable transformation usually preceded by a rigid or affine one Thin-plate splines (SplineKernelTransform) is a deformable transformation belonging to the class of kernel-based transformations that is a composition of and affine and a non-rigid part === Metric === The similarity metric is the mathematical function whose parameters should be optimized to reach the desired registration, and, during the process, it is computed multiple times. Below are reported the available metrics computed employing the reference and the transformed images and the corresponding elastix class names between brackets. Mean squared difference (AdvancedMeanSquares) to be used for mono-modal applications Normalized correlation coefficient (AdvancedNormalizedCorrelation) to be used for images that have an intensity linear relationship Mutual information (AdvancedMattesMutualInformation) to be used for both mono- and multi-modal applications and optimized to reach better performance compared to the normalized version Normalized mutual information (NormalizedMutualInformation) for both mono- and multi-modal applications Kappa statistic (AdvancedKappaStatistic) to be used only for binary images === Sampler === For the computation of the similarity metrics, it is not always necessary to consider all the voxels and, sometimes, it can be useful to use only a fraction of the voxels of the images, i.e. to reduce the execution time for big input images. Below are reported the available criteria for selecting a fraction of the voxels for the similarity metric computation and the corresponding elastix class names between brackets. Full (Full) to employ all the voxels Grid (Grid) to employ a regular grid defined by the user to downsample the image Random (Random) to randomly select a percentage of voxels defined by the users (all voxels have equal probability to be selected) Random coordinate (RandomCoordinate) like the random criterion, but in this case also off-grid positions can be selected to simplify the optimization process === Interpolator === After the application of the transformation, it may occur that the voxels used for the similarity metric computation are at non-voxel positions, so intensity interpolation should be performed to ensure the correctness of the computed values. Below are reported the implemented interpolators and the corresponding elastix class names between brackets. Nearest neighbor (NearestNeighborInterpolator) exploits little resources, but gives low quality results Linear (LinearInterpolator) is sufficient in general applications N-th order B-spline (BSplineInterpolator) can be used to increase the order N, increasing quality and computation time. N=0 and N=1 indicate the nearest neighbor and linear cases respectively. === Optimizer === The optimizer defines the strategy employed for searching the best transformation parameter to reach the correct registration, and it is commonly an iterative strategy. Below are reported some of the implemented optimization strategies. Gradient descent Robbins-Monro, similar to the gradient descent, but employing an approximation of the cost function derivatives A wider range of optimizers is also available, such as Quasi-Newton or evolutionary strategies. === Other features === The elastix software also offers other features that can be employed to speed up the registration procedure and to provide more advanced algorithms to the end-users. Some examples are the introduction of blur and Gaussian pyramid to reduce data complexity, and multi-image and multi-metric framework to deal with more complex applications. == Applications == Elastix has applications mainly in the medical field, where image registration is fundamental to get comprehensive information regarding the analysed anatomical region. It is widely employed in image-guided surgery, tumour monitoring, and treatment assessment. For example, in radiotherapy planning, image registration allows to correctly deliver the treatment and evaluate the obtained results. Thanks to the wide range of implemented algorithms, the use of the elastix software allows physicians and researchers to test different registration pipelines from the simplest to more complex ones, and to save the best one as a configuration file. This file and the fact that the software is completely open-source makes it easy to reproduce the work, that can help supporting the open science paradigm, and allows fast reuse on different patients data. In image-guided surgery, registration time and accuracy are critical points, considering that, during the registration, the patient is on the operating table, and the imag
Scene statistics
Scene statistics is a discipline within the field of perception. It is concerned with the statistical regularities related to scenes. It is based on the premise that a perceptual system is designed to interpret scenes. Biological perceptual systems have evolved in response to physical properties of natural environments. Therefore natural scenes receive a great deal of attention. Natural scene statistics are useful for defining the behavior of an ideal observer in a natural task, typically by incorporating signal detection theory, information theory or estimation theory. == Within-domain versus across-domain == Geisler (2008) distinguishes between four kinds of domains: (1) Physical environments (2) Images/Scenes (3) Neural responses and (4) Behavior. Within the domain of images/scenes one can study the characteristics of information related to redundancy and efficient coding. Across-domain statistics determine how an autonomous system should make inferences about its environment, process information and control its behavior. To study these statistics it is necessary to sample or register information in multiple domains simultaneously. == Applications == === Prediction of picture and video quality === One of the most successful applications of Natural Scenes Statistics Models has been perceptual picture and video quality prediction. For example, the Visual Information Fidelity (VIF) algorithm, which is used to measure the degree of distortion of pictures and videos, is used extensively by the image and video processing communities to assess perceptual quality. This is often after processing, such as compression, which can degrade the appearance of a visual signal. The premise is that the scene statistics are changed by distortion and that the visual system is sensitive to the changes in the scene statistics. VIF is heavily used in the streaming television industry. Other popular picture quality models that use natural scene statistics include BRISQUE and NIQE, both of which are no-reference since they do not require any reference picture to measure quality against.
VLLM
vLLM is an open-source software framework for inference and serving of large language models and related multimodal models. Originally developed at the University of California, Berkeley's Sky Computing Lab, the project is centered on PagedAttention, a memory-management method for transformer key–value caches, and supports features such as continuous batching, distributed inference, quantization, and OpenAI-compatible APIs. According to a project maintainer, the "v" in vLLM originally referred to "virtual", inspired by virtual memory. == History == vLLM was introduced in 2023 by researchers affiliated with the Sky Computing Lab at UC Berkeley. Its core ideas were described in the 2023 paper Efficient Memory Management for Large Language Model Serving with PagedAttention, which presented the system as a high-throughput and memory-efficient serving engine for large language models. In 2025, the PyTorch Foundation announced that vLLM had become a Foundation-hosted project. PyTorch's project page states that the University of California, Berkeley contributed vLLM to the Linux Foundation in July 2024. In January 2026, TechCrunch reported that the creators of vLLM had launched the startup Inferact to commercialize the project, raising $150 million in seed funding. == Architecture == According to its 2023 paper, vLLM was designed to improve the efficiency of large language model serving by reducing memory waste in the key–value cache used during transformer inference. The paper introduced PagedAttention, an algorithm inspired by virtual memory and paging techniques in operating systems, and described vLLM as using block-level memory management and request scheduling to increase throughput while maintaining similar latency. The project documentation and repository describe support for continuous batching, chunked prefill, speculative decoding, prefix caching, quantization, and multiple forms of distributed inference and serving. PyTorch has described vLLM as a high-throughput, memory-efficient inference and serving engine that supports a range of hardware back ends, including NVIDIA and AMD GPUs, Google TPUs, AWS Trainium, and Intel processors.