ISPConfig is an open source hosting control panel for Linux, licensed under BSD license and developed by the company ISPConfig UG. The ISPConfig project was started in autumn 2005 by Till Brehm from the German company projektfarm GmbH. == Overview == Using the dashboard, administrators have the ability to manage websites, email addresses, MySQL and MariaDB as well as PostgreSQL (since version 3.3) databases, FTP accounts, Shell accounts and DNS records through a web-based interface. The software has 4 login levels: administrator, reseller, client, and email-user, each with a different set of permissions. == Operating Systems == ISPConfig is only available on Linux, with CentOS, Debian, and Ubuntu being among the supported distributions. == Features == The following services and features are supported: Management of a single or multiple servers from one control panel. Web server management for Apache HTTP Server and Nginx. Mail server management (with virtual mail users) with spam and antivirus filter using Postfix (software) and Dovecot (software). DNS server management (BIND, Powerdns). Configuration mirroring and clusters. Administrator, reseller, client and mail-user login. Virtual server management for OpenVZ Servers. Website statistics using Webalizer and AWStats
Autoscaling
Autoscaling, (also written as auto scaling, auto-scaling, or known as automatic scaling), is a method used in cloud computing that dynamically adjusts the amount of computational resources in a server farm - typically measured by the number of active servers - automatically based on the load on the farm. For example, the number of servers running behind a web application may be increased or decreased automatically based on the number of active users on the site. Since such metrics may change dramatically throughout the course of the day, and servers are a limited resource that cost money to run even while idle, there is often an incentive to run "just enough" servers to support the current load while still being able to support sudden and large spikes in activity. Autoscaling is helpful for such needs, as it can reduce the number of active servers when activity is low, and launch new servers when activity is high. Autoscaling is closely related to, and builds upon, the idea of load balancing. == Advantages == Autoscaling offers the following advantages: For companies running their own web server infrastructure, autoscaling typically means allowing some servers to go to sleep during times of low load, saving on electricity costs (as well as water costs if water is being used to cool the machines). For companies using infrastructure hosted in the cloud, autoscaling can mean lower bills, because most cloud providers charge based on total usage rather than maximum capacity. Even for companies that cannot reduce the total compute capacity they run or pay for at any given time, autoscaling can help by allowing the company to run less time-sensitive workloads on machines that get freed up by autoscaling during times of low traffic. Autoscaling solutions, such as the one offered by Amazon Web Services, can also take care of replacing unhealthy instances and therefore protecting somewhat against hardware, network, and application failures. Autoscaling can offer greater uptime and more availability in cases where production workloads are variable and unpredictable. Autoscaling differs from having a fixed daily, weekly, or yearly cycle of server use in that it is responsive to actual usage patterns, and thus reduces the potential downside of having too few or too many servers for the traffic load. For instance, if traffic is usually lower at midnight, then a static scaling solution might schedule some servers to sleep at night, but this might result in downtime on a night where people happen to use the Internet more (for instance, due to a viral news event). Autoscaling, on the other hand, can handle unexpected traffic spikes better. == Terminology == In the list below, we use the terminology used by Amazon Web Services (AWS). However, alternative names are noted and terminology that is specific to the names of Amazon services is not used for the names. == Practice == === Amazon Web Services (AWS) === Amazon Web Services launched the Amazon Elastic Compute Cloud (EC2) service in August 2006, that allowed developers to programmatically create and terminate instances (machines). At the time of initial launch, AWS did not offer autoscaling, but the ability to programmatically create and terminate instances gave developers the flexibility to write their own code for autoscaling. Third-party autoscaling software for AWS began appearing around April 2008. These included tools by Scalr and RightScale. RightScale was used by Animoto, which was able to handle Facebook traffic by adopting autoscaling. On May 18, 2009, Amazon launched its own autoscaling feature along with Elastic Load Balancing, as part of Amazon Elastic Compute Cloud. Autoscaling is now an integral component of Amazon's EC2 offering. Autoscaling on Amazon Web Services is done through a web browser or the command line tool. In May 2016 Autoscaling was also offered in AWS ECS Service. On-demand video provider Netflix documented their use of autoscaling with Amazon Web Services to meet their highly variable consumer needs. They found that aggressive scaling up and delayed and cautious scaling down served their goals of uptime and responsiveness best. In an article for TechCrunch, Zev Laderman, the co-founder and CEO of Newvem, a service that helps optimize AWS cloud infrastructure, recommended that startups use autoscaling in order to keep their Amazon Web Services costs low. Various best practice guides for AWS use suggest using its autoscaling feature even in cases where the load is not variable. That is because autoscaling offers two other advantages: automatic replacement of any instances that become unhealthy for any reason (such as hardware failure, network failure, or application error), and automatic replacement of spot instances that get interrupted for price or capacity reasons, making it more feasible to use spot instances for production purposes. Netflix's internal best practices require every instance to be in an autoscaling group, and its conformity monkey terminates any instance not in an autoscaling group in order to enforce this best practice. === Microsoft's Windows Azure === On June 27, 2013, Microsoft announced that it was adding autoscaling support to its Windows Azure cloud computing platform. Documentation for the feature is available on the Microsoft Developer Network. === Oracle Cloud === Oracle Cloud Platform allows server instances to automatically scale a cluster in or out by defining an auto-scaling rule. These rules are based on CPU and/or memory utilization and determine when to add or remove nodes. === Google Cloud Platform === On November 17, 2014, the Google Compute Engine announced a public beta of its autoscaling feature for use in Google Cloud Platform applications. As of March 2015, the autoscaling tool is still in Beta. === Facebook === In a blog post in August 2014, a Facebook engineer disclosed that the company had started using autoscaling to bring down its energy costs. The blog post reported a 27% decline in energy use for low traffic hours (around midnight) and a 10-15% decline in energy use over the typical 24-hour cycle. === Kubernetes Horizontal Pod Autoscaler === Kubernetes Horizontal Pod Autoscaler automatically scales the number of pods in a replication controller, deployment or replicaset based on observed CPU utilization (or, with beta support, on some other, application-provided metrics) == Alternative autoscaling decision approaches == Autoscaling by default uses reactive decision approach for dealing with traffic scaling: scaling only happens in response to real-time changes in metrics. In some cases, particularly when the changes occur very quickly, this reactive approach to scaling is insufficient. Two other kinds of autoscaling decision approaches are described below. === Scheduled autoscaling approach === This is an approach to autoscaling where changes are made to the minimum size, maximum size, or desired capacity of the autoscaling group at specific times of day. Scheduled scaling is useful, for instance, if there is a known traffic load increase or decrease at specific times of the day, but the change is too sudden for reactive approach based autoscaling to respond fast enough. AWS autoscaling groups support scheduled scaling. === Predictive autoscaling === This approach to autoscaling uses predictive analytics. The idea is to combine recent usage trends with historical usage data as well as other kinds of data to predict usage in the future, and autoscale based on these predictions. For parts of their infrastructure and specific workloads, Netflix found that Scryer, their predictive analytics engine, gave better results than Amazon's reactive autoscaling approach. In particular, it was better for: Identifying huge spikes in demand in the near future and getting capacity ready a little in advance Dealing with large-scale outages, such as failure of entire availability zones and regions Dealing with variable traffic patterns, providing more flexibility on the rate of scaling out or in based on the typical level and rate of change in demand at various times of day On November 20, 2018, AWS announced that predictive scaling would be available as part of its autoscaling offering.
EXAPT
EXAPT (a portmanteau of "Extended Subset of APT") is a production-oriented programming language that allows users to generate NC programs with control information for machining tools and facilitates decision-making for production-related issues that may arise during various machining processes. EXAPT was first developed to address industrial requirements. Through the years, the company created additional software for the manufacturing industry. Today, EXAPT offers a suite of SAAS products and services for the manufacturing industry. The trade name, EXAPT, is most commonly associated with the CAD/CAM-System, production data, and tool management software of the German company EXAPT Systemtechnik GmbH based in Aachen, DE. == General == EXAPT is a modularly built programming system for all NC machining operations as Drilling Turning Milling Turn-Milling Nibbling Flame-, laser-, plasma- and water jet cutting Wire eroding Operations with industrial robots Due to the modular structure, the main product groups, EXAPTcam and EXAPTpdo, are gradually expandable and permit individual software for the manufacturing industry used individually and also in a compound with an existing IT environment. == Functionality == EXAPTcam meets the requirements for NC planning, especially for the cutting operations such as turning, drilling, and milling up to 5-axis simultaneous machining. Thereby new process technologies, tool, and machine concepts are constantly involved. In the NC programming data from different sources such as 3D CAD models, drawings or tables can flow in. The possibilities of NC programming reaches from language-oriented to feature-oriented NC programming. The integrated EXAPT knowledge database and intelligent and scalable automatisms support the user. The EXAPT NC planning also covers the generation of production information as clamping and tool plans, presetting data or time calculations. The realistic simulation possibilities of NC planning and NC control data provide with production reliability. EXAPTpdo (EXAPT ProductionsDataOrganization) provides a neutrally applicable technology platform for the information compound of the NC planning - to the shop floor. This applies to all NC production data that are necessary for the set-up of NC machines, for the provision, presetting, and stocking of manufacturing resources and provided by EXAPTpdo in a central database. Besides classical functions of the tool management system (TMS) as the management of cutting tools, measuring, testing and clamping devices the technology data management and tool lifecycle management (TLM) is also included. System-supported "where-used lists" helps to handle the manufacturing resource cycle by secured requirement determination and requirement fulfillment. Unnecessary transports and unplanned dispositive adjustments are dropped, stocks are reduced, set-up times reduced and the throughput is increased. EXAPTpdo synchronizes involved systems within the value chain. Stock systems, MES systems or ERP systems (e.g. from the purchasing or production areas) do not work in isolation from each other but they interact with each other. EXAPTpdo provides the base to Smart Factory, for more flexibility in production and faster communication. == History == With the foundation of the EXAPT-Verein in 1967 as spin-off of the universities Aachen, Berlin and Stuttgart the further development "EXAPT (EXtended Subset of APT)" of the programming language "APT (Automatically Programmed Tool)" was focused and so the first milestone for the EXAPT history was set. In the same year the system EXAPT 1 for drilling and simple milling tasks became available. 1969 The industrial application of EXAPT 2 for the programming of NC machines with 2-axis linear and path control begins. In the following year, the development of the EXAPT modular system starts. 1972 BASIC-EXAPT is provided for the universal, homogeneous programming of all NC tasks. The support is made by the EXAPT applications consultancy. 1973 EXAPT 1.1 is provided for the programming of straight-cut and continuous-path controlled drilling and milling machines and machining centers. At the Hanover Fair (IHA 73) the interactive access to a mainframe via a time-sharing terminal for the part program entry and correction is presented and starts the replacement of the punch card. 1974 The possibilities for the use of process computers for the NC data transfer are leveled out. EXAPT offers the possibility of the result simulation when using plotters with display of tool paths and tools in assignment to the workpiece. In April 1975, the EXAPT NC Systemtechnik GmbH was founded with the aim, of enabling entry into the NC technique for small and medium-sized companies by a complete product and service program. In the following year, the system portfolio is extended with further system modules and service programs and the provision of postprocessors. 1978 The development activities on the EXAPT module system started in 1970 are completed. Using modern software techniques, the different system parts BASIC-EXAPT, EXAPT 1, EXAPT 1.1, and EXAPT 2 are composed of a total system. System support and applications consultancy become a new working focus. From the beginning to the middle of the 1980s Beside new portable software modules for CAD/CAM applications (e. g. CAPEX, NESTEX, CADEX, CADCPL), the first version of the EXAPT DNC system and extensions of the EXAPT NC programming system for the machining of sculptured surfaces are presented. 1988 EXAPT expands the software product range by systems for tool data management (BMO) and production data management (FDO). EXAPT trains more than 1,300 course participants including company-specific courses. 1992 The first version of the completely new product generation EXAPTplus is presented and the agency in Dresden is opened. 1993 The company name "EXAPT NC Systemtechnik GmbH" is changed to "EXAPT Systemtechnik GmbH." EXAPTplus is presented on PC under Windows NT at the EMO '93. The decentralization of the use of EXAPT systems expands the range of applications. In the following year, EXAPT-DNC is executable under Windows on a customary PC. Special hardware is not needed and so it can be used in compound with the database-supported EXAPT production data management system (FDO). 1995 EXAPTplus is also ready for complex application cases such as machining of tubes at extrusion tools. EXAPT-CADI provides the transfer of 2D CAD data to EXAPTplus. With the new office Gießen the marketing is strengthened. In the following year the EXAPT NC editor is developed for the direct processing of NC control data with tool path display and visualization of the tools. In the course of the market entry of more comfortable 3D CAD systems for the solid modelling of components a detailed evaluation of current systems is made in 1997. It is decided to use SolidWorks as a reference system for the solid-oriented NC planning with EXAPT. 1998 The first solution for the transfer of geometry data between SolidWorks and EXAPTplus is generated. The EXAPT organization systems are (beside SQL) also executable under Oracle now. The use of client server solutions supports the data flow in the production. 1999 AFR functions are provided in connection with EXAPTsolid to support a workpiece modelling for NC. The millennium capability is ensured for all EXAPT systems. AFR is a ground-breaking for the integration of third-party products. 2002 EXAPT-BMG is developed for the generation and visualization of tools with additional functions for the assembly from components. The acquisition of tools with their geometric and technological presentation offers extensive support of the NC planning with EXAPT systems. 2003 EXAPTpdo is available to optimize the process chains in production planning and production execution optimally regarding the increasing requirements of changing production conditions. 2004 Diverse system extensions are made in EXAPTplus, EXAPTsolid, EXAPT NC editor, EXAPTpdo for the complete machining on turning/milling centres with result reliability because of more extensive simulation based on realNC (Tecnomatix), for the use of new complex tool systems and the compound use between ERP systems as SAP and intelligent CNC systems. In the following year, EXAPTpdo is extended for the cross-order set-up optimization and provision of manufacturing re-sources especially for single and small series production with connection to purchase and physical portfolio management. 2006 The EXAPT systems are available for extended use as an information platform for production, the time management, and similar requirements. EXAPTsolid is extended for the feature-oriented milling operation and machine simulation. The NC programming of complex machine tools, e.g. three-turret-turning/milling centers is supported by EXAPT systems, as well as the use of multi-functional tools. 2007 A module for 3-5-axis simultaneous milling machining is presented.
Deep Instinct
Deep Instinct is a cybersecurity company that applies deep learning to cybersecurity. The company implements artificial intelligence to the task of preventing and detecting malware. The company was the recipient of the Technology Pioneer by The World Economic Forum in 2017. Lane Bess has been CEO of the company since 2022. == Overview == In 2015, Deep Instinct was founded by Guy Caspi, Dr. Eli David, and Nadav Maman. The headquarters of the company is located in New York City. In July 2017, NVIDIA became an investor. According to Tom's Hardware, NVIDIA’s investment enabled access to a GPU-based neural network and CUDA platform, which they were using to achieve maximum vulnerability detection rates. As of February 2020, the company had raised $43 million in Series C funding round. In April 2021, Deep Instinct raised $100 million in Series D funding to accelerate growth. == Partnerships == In April 2019, Deep Instinct partnered with Chinese artist, Guo O. Dong on an art project titled, The Persistence of Chaos, consisting of a laptop infected with 6 pieces of malware that represented $95 billion in damages. The art was auctioned with a final bid of $1,345,000. In the same year, Globes reported that, HP Inc partnered with Deep Instinct to launch their security solution HP SureSense, which has been applied to the EliteBook and Zbook devices.
Deep learning
In machine learning, deep learning (DL) focuses on utilizing multilayered neural networks to perform tasks such as classification, regression, and representation learning. The field takes inspiration from biological neuroscience and revolves around stacking artificial neurons into layers and "training" them to process data. The adjective "deep" refers to the use of multiple layers (ranging from three to several hundred or thousands) in the network. Methods used can be supervised, semi-supervised or unsupervised. Some common deep learning network architectures include fully connected networks, deep belief networks, recurrent neural networks, convolutional neural networks, generative adversarial networks, transformers, and neural radiance fields. These architectures have been applied to fields including computer vision, speech recognition, natural language processing, machine translation, bioinformatics, drug design, medical image analysis, climate science, material inspection and board game programs, where they have produced results comparable to and in some cases surpassing human expert performance. Early forms of neural networks were inspired by information processing and distributed communication nodes in biological systems, particularly the human brain. However, current neural networks do not intend to model the brain function of organisms, and are generally seen as low-quality models for that purpose. == Overview == Most modern deep learning models are based on multi-layered neural networks such as convolutional neural networks and transformers, although they can also include propositional formulas or latent variables organized layer-wise in deep generative models such as the nodes in deep belief networks and deep Boltzmann machines. Fundamentally, deep learning refers to a class of machine learning algorithms in which a hierarchy of layers is used to transform input data into a progressively more abstract and composite representation. For example, in an image recognition model, the raw input may be an image (represented as a tensor of pixels). The first representational layer may attempt to identify basic shapes such as lines and circles, the second layer may compose and encode arrangements of edges, the third layer may encode a nose and eyes, and the fourth layer may recognize that the image contains a face. Importantly, a deep learning process can learn which features to optimally place at which level on its own. Prior to deep learning, machine learning techniques often involved hand-crafted feature engineering to transform the data into a more suitable representation for a classification algorithm to operate on. In the deep learning approach, features are not hand-crafted and the model discovers useful feature representations from the data automatically. This does not eliminate the need for hand-tuning; for example, varying numbers of layers and layer sizes can provide different degrees of abstraction. The word "deep" in "deep learning" refers to the number of layers through which the data is transformed. More precisely, deep learning systems have a substantial credit assignment path (CAP) depth. The CAP is the chain of transformations from input to output. CAPs describe potentially causal connections between input and output. For a feedforward neural network, the depth of the CAPs is that of the network and is the number of hidden layers plus one (as the output layer is also parameterized). For recurrent neural networks, in which a signal may propagate through a layer more than once, the CAP depth is potentially unlimited. No universally agreed-upon threshold of depth divides shallow learning from deep learning, but most researchers agree that deep learning involves CAP depth higher than two. CAP of depth two has been shown to be a universal approximator in the sense that it can emulate any function. Beyond that, more layers do not add to the function approximator ability of the network. Deep models (CAP > two) are able to extract better features than shallow models and hence, extra layers help in learning the features effectively. Deep learning architectures can be constructed with a greedy layer-by-layer method. Deep learning helps to disentangle these abstractions and pick out which features improve performance. Deep learning algorithms can be applied to unsupervised learning tasks. This is an important benefit because unlabeled data is more abundant than labeled data. Examples of deep structures that can be trained in an unsupervised manner are deep belief networks. The term deep learning was introduced to the machine learning community by Rina Dechter in 1986, and to artificial neural networks by Igor Aizenberg and colleagues in 2000, in the context of Boolean threshold neurons. The etymology of the term is more complicated. == Interpretations == Deep neural networks are generally interpreted in terms of the universal approximation theorem or probabilistic inference. The classic universal approximation theorem concerns the capacity of feedforward neural networks with a single hidden layer of finite size to approximate continuous functions. In 1989, the first proof was published by George Cybenko for sigmoid activation functions and was generalised to feed-forward multi-layer architectures in 1991 by Kurt Hornik. Recent work also showed that universal approximation also holds for non-bounded activation functions such as Kunihiko Fukushima's rectified linear unit. The universal approximation theorem for deep neural networks concerns the capacity of networks with bounded width but the depth is allowed to grow. Lu et al. proved that if the width of a deep neural network with ReLU activation is strictly larger than the input dimension, then the network can approximate any Lebesgue integrable function; if the width is smaller or equal to the input dimension, then a deep neural network is not a universal approximator. The probabilistic interpretation derives from the field of machine learning. It features inference, as well as the optimization concepts of training and testing, related to fitting and generalization, respectively. More specifically, the probabilistic interpretation considers the activation nonlinearity as a cumulative distribution function. The probabilistic interpretation led to the introduction of dropout as regularizer in neural networks. The probabilistic interpretation was introduced by researchers including Hopfield, Widrow and Narendra and popularized in surveys such as the one by Bishop. == History == === Before 1980 === There are two types of artificial neural network (ANN): feedforward neural network (FNN) or multilayer perceptron (MLP) and recurrent neural networks (RNN). RNNs have cycles in their connectivity structure, whereas FNNs do not. In the 1920s, Wilhelm Lenz and Ernst Ising created the Ising model which is essentially a non-learning RNN architecture consisting of neuron-like threshold elements. In 1972, Shun'ichi Amari made this architecture adaptive. His learning RNN was republished by John Hopfield in 1982. Other early recurrent neural networks were published by Kaoru Nakano in 1971. Already in 1948, Alan Turing produced work on "Intelligent Machinery" that was not published in his lifetime, containing "ideas related to artificial evolution and learning RNNs". Frank Rosenblatt (1958) proposed the perceptron, an MLP with 3 layers: an input layer, a hidden layer with randomized weights that did not learn, and an output layer. He later published a 1962 book that also introduced variants and computer experiments, including a version with four-layer perceptrons "with adaptive preterminal networks" where the last two layers have learned weights (here he credits H. D. Block and B. W. Knight). The book cites an earlier network by R. D. Joseph (1960) "functionally equivalent to a variation of" this four-layer system (the book mentions Joseph over 30 times). Should Joseph therefore be considered the originator of proper adaptive multilayer perceptrons with learning hidden units? Unfortunately, the learning algorithm was not a functional one, and fell into oblivion. The first working deep learning algorithm was the Group method of data handling, a method to train arbitrarily deep neural networks, published by Alexey Ivakhnenko and Lapa in 1965. They regarded it as a form of polynomial regression, or a generalization of Rosenblatt's perceptron to handle more complex, nonlinear, and hierarchical relationships. A 1971 paper described a deep network with eight layers trained by this method, which is based on layer by layer training through regression analysis. Superfluous hidden units are pruned using a separate validation set. Since the activation functions of the nodes are Kolmogorov-Gabor polynomials, these were also the first deep networks with multiplicative units or "gates". The first deep learning multilayer perceptron trained by stochastic gradient descent was published in 1967 by Shun'ichi
Sentence embedding
In natural language processing, a sentence embedding is a representation of a sentence as a vector of numbers which encodes meaningful semantic information. State of the art embeddings are based on the learned hidden layer representation of dedicated sentence transformer models. BERT pioneered an approach involving the use of a dedicated [CLS] token prepended to the beginning of each sentence inputted into the model; the final hidden state vector of this token encodes information about the sentence and can be fine-tuned for use in sentence classification tasks. In practice however, BERT's sentence embedding with the [CLS] token achieves poor performance, often worse than simply averaging non-contextual word embeddings. SBERT later achieved superior sentence embedding performance by fine tuning BERT's [CLS] token embeddings through the usage of a siamese neural network architecture on the SNLI dataset. Other approaches are loosely based on the idea of distributional semantics applied to sentences. Skip-Thought trains an encoder-decoder structure for the task of neighboring sentences predictions; this has been shown to achieve worse performance than approaches such as InferSent or SBERT. An alternative direction is to aggregate word embeddings, such as those returned by Word2vec, into sentence embeddings. The most straightforward approach is to simply compute the average of word vectors, known as continuous bag-of-words (CBOW). However, more elaborate solutions based on word vector quantization have also been proposed. One such approach is the vector of locally aggregated word embeddings (VLAWE), which demonstrated performance improvements in downstream text classification tasks. == Applications == In recent years, sentence embedding has seen a growing level of interest due to its applications in natural language queryable knowledge bases through the usage of vector indexing for semantic search. LangChain for instance utilizes sentence transformers for purposes of indexing documents. In particular, an indexing is generated by generating embeddings for chunks of documents and storing (document chunk, embedding) tuples. Then given a query in natural language, the embedding for the query can be generated. A top k similarity search algorithm is then used between the query embedding and the document chunk embeddings to retrieve the most relevant document chunks as context information for question answering tasks. This approach is also known formally as retrieval-augmented generation. Though not as predominant as BERTScore, sentence embeddings are commonly used for sentence similarity evaluation which sees common use for the task of optimizing a Large language model's generation parameters is often performed via comparing candidate sentences against reference sentences. By using the cosine-similarity of the sentence embeddings of candidate and reference sentences as the evaluation function, a grid-search algorithm can be utilized to automate hyperparameter optimization. == Evaluation == A way of testing sentence encodings is to apply them on Sentences Involving Compositional Knowledge (SICK) corpus for both entailment (SICK-E) and relatedness (SICK-R). In the best results are obtained using a BiLSTM network trained on the Stanford Natural Language Inference (SNLI) Corpus. The Pearson correlation coefficient for SICK-R is 0.885 and the result for SICK-E is 86.3. A slight improvement over previous scores is presented in: SICK-R: 0.888 and SICK-E: 87.8 using a concatenation of bidirectional Gated recurrent unit.
Retrieval-augmented generation
Retrieval-augmented generation (RAG) is a technique that enables large language models (LLMs) to retrieve and incorporate new information from external data sources. With RAG, LLMs first refer to a specified set of documents, then respond to user queries. These documents supplement information from the LLM's pre-existing training data. This allows LLMs to use domain-specific and/or updated information that is not available in the training data. For example, this enables LLM-based chatbots to access internal company data or generate responses based on authoritative sources. RAG improves LLMs by incorporating information retrieval before generating responses. Unlike LLMs that rely on static training data, RAG pulls relevant text from databases, uploaded documents, or web sources. According to Ars Technica, "RAG is a way of improving LLM performance, in essence by blending the LLM process with a web search or other document look-up process to help LLMs stick to the facts." This method helps reduce AI hallucinations, which have caused chatbots to describe policies that don't exist, or recommend nonexistent legal cases to lawyers that are looking for citations to support their arguments. RAG also reduces the need to retrain LLMs with new data, saving on computational and financial costs. Beyond efficiency gains, RAG also allows LLMs to include sources in their responses, so users can verify the cited sources. This provides greater transparency, as users can cross-check retrieved content to ensure accuracy and relevance. The term retrieval-augmented generation (RAG) was introduced in a 2020 paper that described combining a parametric language model with a non-parametric external memory accessed through retrieval at inference time. == RAG and LLM limitations == LLMs can provide incorrect information. For example, when Google first demonstrated its LLM tool "Google Bard" (later re-branded to Gemini), the LLM provided incorrect information about the James Webb Space Telescope. This error contributed to a $100 billion decline in Google's stock value. RAG is used to prevent these errors, but it does not solve all the problems. For example, LLMs can generate misinformation even when pulling from factually correct sources if they misinterpret the context. MIT Technology Review gives the example of an AI-generated response stating, "The United States has had one Muslim president, Barack Hussein Obama." The model retrieved this from an academic book rhetorically titled Barack Hussein Obama: America's First Muslim President? The LLM did not "know" or "understand" the context of the title, generating a false statement. LLMs with RAG are programmed to prioritize new information. This technique has been called "prompt stuffing." Without prompt stuffing, the LLM's input is generated by a user; with prompt stuffing, additional relevant context is added to this input to guide the model's response. This approach provides the LLM with key information early in the prompt, encouraging it to prioritize the supplied data over pre-existing training knowledge. == Process == Retrieval-augmented generation (RAG) enhances large language models (LLMs) by incorporating an information-retrieval mechanism that allows models to access and utilize additional data beyond their original training set. Ars Technica notes that "when new information becomes available, rather than having to retrain the model, all that's needed is to augment the model's external knowledge base with the updated information" ("augmentation"). IBM states that "in the generative phase, the LLM draws from the augmented prompt and its internal representation of its training data to synthesize" an answer. === RAG key stages === Typically, the data to be referenced is converted into LLM embeddings, numerical representations in the form of a large vector space. RAG can be used on unstructured (usually text), semi-structured, or structured data (for example knowledge graphs). These embeddings are then stored in a vector database to allow for document retrieval. Given a user query, a document retriever is first called to select the most relevant documents that will be used to augment the query. This comparison can be done using a variety of methods, which depend in part on the type of indexing used. The model feeds this relevant retrieved information into the LLM via prompt engineering of the user's original query. Newer implementations (as of 2023) can also incorporate specific augmentation modules with abilities such as expanding queries into multiple domains and using memory and self-improvement to learn from previous retrievals. Finally, the LLM can generate output based on both the query and the retrieved documents. Some models incorporate extra steps to improve output, such as the re-ranking of retrieved information, context selection, and fine-tuning. == Applications == Retrieval-augmented generation is used in applications where generated responses need to be grounded in external or frequently updated information. Commonly cited use cases include search engines, question-answering systems, customer support chatbots, enterprise knowledge assistants, content generation, recommendation systems, retail and e-commerce, and industrial or manufacturing workflows. In healthcare, RAG has been studied as a way to ground large language model outputs in external medical knowledge sources, although reviews have noted continuing challenges around evaluation, ethics, and clinical reliability. == Improvements == Improvements to the basic process above can be applied at different stages in the RAG flow. === Encoder === These methods focus on the encoding of text as either dense or sparse vectors. Sparse vectors, which encode the identity of a word, are typically dictionary-length and contain mostly zeros. Dense vectors, which encode meaning, are more compact and contain fewer zeros. Various enhancements can improve the way similarities are calculated in the vector stores (databases). Performance improves by optimizing how vector similarities are calculated. Dot products enhance similarity scoring, while approximate nearest neighbor (ANN) searches improve retrieval efficiency over K-nearest neighbors (KNN) searches. Accuracy may be improved with Late Interactions, which allow the system to compare words more precisely after retrieval. This helps refine document ranking and improve search relevance. Hybrid vector approaches may be used to combine dense vector representations with sparse one-hot vectors, taking advantage of the computational efficiency of sparse dot products over dense vector operations. Other retrieval techniques focus on improving accuracy by refining how documents are selected. Some retrieval methods combine sparse representations, such as SPLADE, with query expansion strategies to improve search accuracy and recall. === Retriever-centric methods === These methods aim to enhance the quality of document retrieval in vector databases: Pre-training the retriever using the Inverse Cloze Task (ICT), a technique that helps the model learn retrieval patterns by predicting masked text within documents. Supervised retriever optimization aligns retrieval probabilities with the generator model's likelihood distribution. This involves retrieving the top-k vectors for a given prompt, scoring the generated response's perplexity, and minimizing KL divergence between the retriever's selections and the model's likelihoods to refine retrieval. Reranking techniques can refine retriever performance by prioritizing the most relevant retrieved documents during training. === Language model === By redesigning the language model with the retriever in mind, a 25-time smaller network can get comparable perplexity as its much larger counterparts. Because it is trained from scratch, this method (Retro) incurs the high cost of training runs that the original RAG scheme avoided. The hypothesis is that by giving domain knowledge during training, Retro needs less focus on the domain and can devote its smaller weight resources only to language semantics. The redesigned language model is shown here. It has been reported that Retro is not reproducible, so modifications were made to make it so. The more reproducible version is called Retro++ and includes in-context RAG. === Chunking === Chunking involves various strategies for breaking up the data into vectors so the retriever can find details in it. Three types of chunking strategies are: Fixed length with overlap. This is fast and easy. Overlapping consecutive chunks helps to maintain semantic context across chunks. Syntax-based chunks can break the document up into sentences. Libraries such as spaCy or NLTK can also help. File format-based chunking. Certain file types have natural chunks built in, and it's best to respect them. For example, code files are best chunked and vectorized as whole functions or classes. HTML files should leave