A large language model (LLM) is a neural network trained on a vast amount of text for natural language processing tasks, especially language generation. LLMs can typically generate, summarize, translate and analyze text in many contexts, and are a foundational technology behind modern chatbots. Biased or inaccurate training data can make an LLM's output less reliable. As of 2026, the most capable LLMs are based on transformer architectures, which, according to the 2017 paper "Attention Is All You Need", can be more efficient and parallelizable than earlier statistical and recurrent neural network models. Benchmark evaluations for LLMs attempt to measure model reasoning, factual accuracy, alignment, and safety. == History == Before the emergence of transformer-based models in 2017, some language models were considered large relative to the computational and data constraints of their time. In the early 1990s, IBM's statistical models pioneered word alignment techniques for machine translation, laying the groundwork for corpus-based language modeling. In 2001, a smoothed n-gram model, such as those employing Kneser–Ney smoothing, trained on 300 million words, achieved state-of-the-art perplexity on benchmark tests. During the 2000s, with the rise of widespread internet access, researchers began compiling massive text datasets from the web ("web as corpus") to train statistical language models. Moving beyond n-gram models, researchers started in 2000 to use neural networks as language models. Following the breakthrough of deep neural networks in image classification around 2012, similar architectures were adapted for language tasks. This shift was marked by the development of word embeddings (e.g., Word2Vec by Mikolov in 2013) and sequence-to-sequence (seq2seq) models using LSTM. In 2016, Google transitioned its translation service to neural machine translation (NMT), replacing statistical phrase-based models with deep recurrent neural networks. These early NMT systems used LSTM-based encoder-decoder architectures, as they preceded the invention of transformers. At the 2017 NeurIPS conference, Google researchers introduced the transformer architecture in their landmark paper "Attention Is All You Need". This paper's goal was to improve upon 2014 seq2seq technology, and was based mainly on the attention mechanism developed by Bahdanau et al. in 2014. The following year in 2018, BERT was introduced and quickly became "ubiquitous". Though the original transformer has both encoder and decoder blocks, BERT is an encoder-only model. Academic and research usage of BERT began to decline in 2023, following rapid improvements in the abilities of decoder-only models (such as GPT) to solve tasks via prompting. Although decoder-only GPT-1 was introduced in 2018, it was GPT-2 in 2019 that caught widespread attention because OpenAI claimed to have initially deemed it too powerful to release publicly, out of fear of malicious use. GPT-3 in 2020 went a step further and as of 2025 is available only via API with no offering of downloading the model to execute locally. But it was the consumer-facing chatbot ChatGPT in late 2022 that received extensive media coverage and public attention by 2023. The 2023 GPT-4 was praised for its increased accuracy and as a "holy grail" for its multimodal capabilities. OpenAI did not reveal the high-level architecture and the number of parameters of GPT-4. The release of ChatGPT led to an uptick in LLM usage across several research subfields of computer science, including robotics, software engineering, and societal impact work. In 2024, OpenAI released the reasoning model OpenAI o1, which generates long chains of thought before returning a final answer. Many LLMs with parameter counts comparable to those of OpenAI's GPT series have been developed. Since 2022, weights-available models have been gaining popularity, especially at first with BLOOM and LLaMA, though both have restrictions on usage and deployment. Mistral AI's open-weight models Mistral 7B and Mixtral 8x7B have a more permissive Apache License. In January 2025, DeepSeek released DeepSeek R1, a 671-billion-parameter open-weight model that performs comparably to OpenAI o1 but at a much lower price per token for users. Since 2023, many LLMs have been trained to be multimodal, having the ability to also process or generate other types of data, such as images, audio, or 3D meshes. Open-weight LLMs have become more influential since 2023. Per Vake et al. (2025), community-driven contributions to open-weight models improve their efficiency and performance via collaborative platforms such as Hugging Face. == Dataset preprocessing == === Tokenization === As machine learning algorithms process numbers rather than text, the text must be converted to numbers. In the first step, a vocabulary is decided upon, then integer indices are arbitrarily but uniquely assigned to each vocabulary entry, and finally, an embedding is associated with the integer index. Algorithms include byte-pair encoding (BPE) and WordPiece. There are also special tokens serving as control characters, such as [MASK] for masked-out token (as used in BERT), and [UNK] ("unknown") for characters not appearing in the vocabulary. Also, some special symbols are used to denote special text formatting. For example, "Ġ" denotes a preceding whitespace in RoBERTa and GPT and "##" denotes continuation of a preceding word in BERT. For example, the BPE tokenizer used by the legacy version of GPT-3 would split tokenizer: texts -> series of numerical "tokens" as Tokenization also compresses the datasets. Because LLMs generally require input to be an array that is not jagged, the shorter texts must be "padded" until they match the length of the longest one. ==== Byte-pair encoding ==== As an example, consider a tokenizer based on byte-pair encoding. In the first step, all unique characters (including blanks and punctuation marks) are treated as an initial set of n-grams (i.e. initial set of uni-grams). Successively the most frequent pair of adjacent characters is merged into a bi-gram and all instances of the pair are replaced by it. All occurrences of adjacent pairs of (previously merged) n-grams that most frequently occur together are then again merged into even lengthier n-gram, until a vocabulary of prescribed size is obtained. After a tokenizer is trained, any text can be tokenized by it, as long as it does not contain characters not appearing in the initial-set of uni-grams. === Dataset cleaning === In the context of training LLMs, datasets are typically cleaned by removing low-quality, duplicated, or toxic data. Cleaned datasets can increase training efficiency and lead to improved downstream performance. A trained LLM can be used to clean datasets for training a further LLM. With the increasing proportion of LLM-generated content on the web, data cleaning in the future may include filtering out such content. LLM-generated content can pose a problem if the content is similar to human text (making filtering difficult) but of lower quality (degrading performance of models trained on it). === Synthetic data === Training of largest language models might need more linguistic data than naturally available, or that the naturally occurring data is of insufficient quality. In these cases, synthetic data might be used. == Training == An LLM is a type of foundation model (large X model) trained on language. LLMs can be trained in different ways. In particular, GPT models are first pretrained to predict the next word on a large amount of data, before being fine-tuned. === Cost === Substantial infrastructure is necessary for training the largest models. The tendency towards larger models is visible in the list of large language models. For example, the training of GPT-2 (i.e. a 1.5-billion-parameter model) in 2019 cost $50,000, while training of the PaLM (i.e. a 540-billion-parameter model) in 2022 cost $8 million, and Megatron-Turing NLG 530B (in 2021) cost around $11 million. The qualifier "large" in "large language model" is inherently vague, as there is no definitive threshold for the number of parameters required to qualify as "large". === Fine-tuning === Before being fine-tuned, most LLMs are next-token predictors. The fine-tuning shapes the LLM's behavior via techniques like reinforcement learning from human feedback (RLHF) or constitutional AI. Instruction fine-tuning is a form of supervised learning used to teach LLMs to follow user instructions. In 2022, OpenAI demonstrated InstructGPT, a version of GPT-3 similarly fine-tuned to follow instructions. Reinforcement learning from human feedback (RLHF) involves training a reward model to predict which text humans prefer. Then, the LLM can be fine-tuned through reinforcement learning to better satisfy this reward model. Since humans typically prefer truthful, helpful and harmless answers, RLHF favors such answers. == Architecture == LLMs are generally based on the tra
Spanner (database)
Spanner is a distributed SQL database management and storage service developed by Google. It provides features such as global transactions, strongly consistent reads, and automatic multi-site replication and failover. Spanner is used in Google F1, the database for its advertising business Google Ads, as well as Gmail and Google Photos. == Features == Spanner stores large amounts of mutable structured data. Spanner allows users to perform arbitrary queries using SQL with relational data while maintaining strong consistency and high availability for that data with synchronous replication. Key features of Spanner: Transactions can be applied across rows, columns, tables, and databases within a Spanner universe. Clients can control the replication and placement of data using automatic multi-site replication and failover. Replication is synchronous and strongly consistent. Reads are strongly consistent and data is versioned to allow for stale reads: clients can read previous versions of data, subject to garbage collection windows. Supports a native SQL interface for reading and writing data. Support for Graph Query Language == History == Spanner was first described in 2012 for internal Google data centers. Spanner's SQL capability was added in 2017 and documented in a SIGMOD 2017 paper. It became available as part of Google Cloud Platform in 2017, under the name "Cloud Spanner". == Architecture == Spanner uses the Paxos algorithm as part of its operation to shard (partition) data across up to hundreds of servers. It makes heavy use of hardware-assisted clock synchronization using GPS clocks and atomic clocks to ensure global consistency. TrueTime is the brand name for Google's distributed cloud infrastructure, which provides Spanner with the ability to generate monotonically increasing timestamps in data centers around the world. Google's F1 SQL database management system (DBMS) is built on top of Spanner, replacing Google's custom MySQL variant.
Artificial intelligence industry in China
The roots of the development of artificial intelligence in the People's Republic of China started in the late 1970s following Deng Xiaoping's reform and opening up emphasizing science and technology as the country's primary productive force. The initial stages of China's AI development were slow and encountered significant challenges due to lack of resources and talent. At the beginning China was behind most Western countries in terms of AI development. A majority of the research was led by scientists who had received higher education abroad. Since 2006, the Chinese government has steadily developed a national agenda for artificial intelligence development and emerged as one of the leading nations in artificial intelligence research and development. In 2016, the Chinese Communist Party (CCP) released its 13th Five-Year Plan in which it aimed to become a global AI leader by 2030. As of 2025, China is considered to be a world leader in AI technology along with the United States. The State Council has a list of "national AI teams" including fifteen China-based companies, including Baidu, Tencent, Alibaba, SenseTime, and iFlytek. Each company should lead the development of a designated specialized AI sector in China, such as facial recognition, software/hardware, and speech recognition. China's rapid AI development has significantly impacted Chinese society in many areas, including the socio-economic, military, intelligence, and political spheres. Agriculture, transportation, accommodation and food services, and manufacturing are the top industries that would be the most impacted by further AI deployment. The private sector, university laboratories, and the military are working collaboratively in many aspects as there are few current existing boundaries. In 2021, China published the Data Security Law of the People's Republic of China, its first national law addressing AI-related ethical concerns. In October 2022, the United States federal government announced a series of export controls and trade restrictions intended to restrict China's access to advanced computer chips for AI applications. In 2023, the Cyberspace Administration of China issued guidelines requiring that AI content upholds the ideology of the CCP including Core Socialist Values, avoids discrimination, respects intellectual property rights, and safeguards user data. In 2025, the Chinese government issued a document regarding training data, requiring companies to use as little as data deemed "unsafe" as possible, as well as requiring companies to test models regularly. Concerns have been raised about the effects of the Chinese government's censorship regime on the development of generative artificial intelligence and long-term talent acquisition with state of the country's demographics. Others have noted that official notions of AI safety require following the priorities of the CCP and are antithetical to standards in democratic societies and raised concerns about the extension of China's system of mass surveillance and censorship abroad. == History == The Chinese term for artificial intelligence (réngōngzhìnéng 人工智能) connotes "humanmade" intelligence. The term developed as mid-20th century localisation of the Japanese term jinko chino. The research and development of artificial intelligence in China started in the 1980s, with the announcement by Deng Xiaoping of the importance of science and technology for China's economic growth. === Late 1970s to early 2010s === Chinese artificial intelligence research and development began in late 1970s after Deng Xiaoping's reform and opening up. China's first national conference on AI occurred in 1979. Academic journals in the late 1970s began publishing literature reviews of Western research on AI topics. In the 1980s, a group of Chinese scientists launched AI research led by Qian Xuesen and Wu Wenjun. However, during the time, China's society still had a generally conservative view towards AI. In the early 1980s, Science Press published translated versions of Western textbooks such as Patrick Winston's Artificial Intelligence and Nils John Nilsson's Principles of Artificial Intelligence. In 1980, a journal of the Chinese Academy of Sciences convened its first annual National Symposium on Artificial Intelligence, which included national and international scholars like Herbert A. Simon. The Chinese Association for Artificial Intelligence (CAAI) was founded in September 1981 and was authorized by the Ministry of Civil Affairs. CAAI has continued to be the largest AI association in China as of 2025. In 1982, CAAI began publishing the Artificial Intelligence Journal, which published early AI research by Chinese academics. In the 1980s, Chinese research on AI was influenced by the field of cybernetics, particularly the work of Norbert Weiner and his text Cybernetics: Or Control and Communication in the Animal and the Machine. Chinese researchers at the time sought to situate AI as part of a broader "Intelligence Science" field which would include disciplines like mathematics, computer science, cognitive science, social sciences, and philosophy. In 1987, Tsinghua University began a research publication on AI. Beginning in 1993, smart automation and intelligence have been part of China's national technology plan. Since the 2000s, the Chinese government has further expanded its research and development funds for AI and the number of government-sponsored research projects has dramatically increased. In 2006, China announced a policy priority for the development of artificial intelligence, which was included in the National Medium and Long Term Plan for the Development of Science and Technology (2006–2020), released by the State Council. In the same year, artificial intelligence was also mentioned in the 11th Five-Year Plan. In 2011, the Association for the Advancement of Artificial Intelligence (AAAI) established a branch in Beijing, China. At same year, the Wu Wenjun Artificial Intelligence Science and Technology Award was founded in honor of Chinese mathematician Wu Wenjun, and it became the highest award for Chinese achievements in the field of artificial intelligence. The first award ceremony was held on May 14, 2012. In 2013, the International Joint Conferences on Artificial Intelligence (IJCAI) was held in Beijing, marking the first time the conference was held in China. This event coincided with the Chinese government's announcement of the "Chinese Intelligence Year," a significant milestone in China's development of artificial intelligence. === Late 2010s to early 2020s === AI became a major issue of commercial, public, and political focus in China in the latter half of the 2010s. Various interpretations of the primary cause for this increased focus exist, with some analyses focusing on the 2016 Go match between Google's AlphaGo and Lee Sedol, others emphasising the U.S. increasing trade restrictions on China's technology industries and the desire to achieve national technological self-sufficiency. The State Council of China issued "A Next Generation Artificial Intelligence Development Plan" (State Council Document [2017] No. 35) on 20 July 2017. In the document, the CCP Central Committee and the State Council urged governing bodies in China to promote the development of artificial intelligence. Specifically, the plan described AI as a strategic technology that has become a "focus of international competition".:2 The document urged significant investment in a number of strategic areas related to AI and called for close cooperation between the state and private sectors. It set the goal of China becoming the preeminent country for AI research and application by 2030. During the general secretaryship of Xi Jinping, artificial intelligence has been a focus of the CCP's military-civil fusion efforts. On the occasion of Xi's speech at the first plenary meeting of the Central Military-Civil Fusion Development Committee (CMCFDC), scholars from the National Defense University wrote in the PLA Daily that the "transferability of social resources" between economic and military ends is an essential component to being a great power. During the Two Sessions 2017,"artificial intelligence plus" was proposed to be elevated to a strategic level. The same year witnessed the emergence of multiple application-level usages in the medical field according to reports. In 2018, Xinhua News Agency, in partnership with Tencent's subsidiary Sogou, launched its first artificial intelligence-generated news anchor. In 2018, the State Council budgeted $2.1 billion for an AI industrial park in Mentougou district. In order to achieve this the State Council stated the need for massive talent acquisition, theoretical and practical developments, as well as public and private investments. Some of the stated motivations that the State Council gave for pursuing its AI strategy include the potential of artificial intelligence for industrial transformation, better social
Predictor–corrector method
In numerical analysis, predictor–corrector methods belong to a class of algorithms designed to integrate ordinary differential equations – to find an unknown function that satisfies a given differential equation. All such algorithms proceed in two steps: The initial, "prediction" step, starts from a function fitted to the function-values and derivative-values at a preceding set of points to extrapolate ("anticipate") this function's value at a subsequent, new point. The next, "corrector" step refines the initial approximation by using the predicted value of the function and another method to interpolate that unknown function's value at the same subsequent point. == Predictor–corrector methods for solving ODEs == When considering the numerical solution of ordinary differential equations (ODEs), a predictor–corrector method typically uses an explicit method for the predictor step and an implicit method for the corrector step. === Example: Euler method with the trapezoidal rule === A simple predictor–corrector method (known as Heun's method) can be constructed from the Euler method (an explicit method) and the trapezoidal rule (an implicit method). Consider the differential equation y ′ = f ( t , y ) , y ( t 0 ) = y 0 , {\displaystyle y'=f(t,y),\quad y(t_{0})=y_{0},} and denote the step size by h {\displaystyle h} . First, the predictor step: starting from the current value y i {\displaystyle y_{i}} , calculate an initial guess value y ~ i + 1 {\displaystyle {\tilde {y}}_{i+1}} via the Euler method, y ~ i + 1 = y i + h f ( t i , y i ) . {\displaystyle {\tilde {y}}_{i+1}=y_{i}+hf(t_{i},y_{i}).} Next, the corrector step: improve the initial guess using trapezoidal rule, y i + 1 = y i + 1 2 h ( f ( t i , y i ) + f ( t i + 1 , y ~ i + 1 ) ) . {\displaystyle y_{i+1}=y_{i}+{\tfrac {1}{2}}h{\bigl (}f(t_{i},y_{i})+f(t_{i+1},{\tilde {y}}_{i+1}){\bigr )}.} That value is used as the next step. === PEC mode and PECE mode === There are different variants of a predictor–corrector method, depending on how often the corrector method is applied. The Predict–Evaluate–Correct–Evaluate (PECE) mode refers to the variant in the above example: y ~ i + 1 = y i + h f ( t i , y i ) , y i + 1 = y i + 1 2 h ( f ( t i , y i ) + f ( t i + 1 , y ~ i + 1 ) ) . {\displaystyle {\begin{aligned}{\tilde {y}}_{i+1}&=y_{i}+hf(t_{i},y_{i}),\\y_{i+1}&=y_{i}+{\tfrac {1}{2}}h{\bigl (}f(t_{i},y_{i})+f(t_{i+1},{\tilde {y}}_{i+1}){\bigr )}.\end{aligned}}} It is also possible to evaluate the function f only once per step by using the method in Predict–Evaluate–Correct (PEC) mode: y ~ i + 1 = y i + h f ( t i , y ~ i ) , y i + 1 = y i + 1 2 h ( f ( t i , y ~ i ) + f ( t i + 1 , y ~ i + 1 ) ) . {\displaystyle {\begin{aligned}{\tilde {y}}_{i+1}&=y_{i}+hf(t_{i},{\tilde {y}}_{i}),\\y_{i+1}&=y_{i}+{\tfrac {1}{2}}h{\bigl (}f(t_{i},{\tilde {y}}_{i})+f(t_{i+1},{\tilde {y}}_{i+1}){\bigr )}.\end{aligned}}} Additionally, the corrector step can be repeated in the hope that this achieves an even better approximation to the true solution. If the corrector method is run twice, this yields the PECECE mode: y ~ i + 1 = y i + h f ( t i , y i ) , y ^ i + 1 = y i + 1 2 h ( f ( t i , y i ) + f ( t i + 1 , y ~ i + 1 ) ) , y i + 1 = y i + 1 2 h ( f ( t i , y i ) + f ( t i + 1 , y ^ i + 1 ) ) . {\displaystyle {\begin{aligned}{\tilde {y}}_{i+1}&=y_{i}+hf(t_{i},y_{i}),\\{\hat {y}}_{i+1}&=y_{i}+{\tfrac {1}{2}}h{\bigl (}f(t_{i},y_{i})+f(t_{i+1},{\tilde {y}}_{i+1}){\bigr )},\\y_{i+1}&=y_{i}+{\tfrac {1}{2}}h{\bigl (}f(t_{i},y_{i})+f(t_{i+1},{\hat {y}}_{i+1}){\bigr )}.\end{aligned}}} The PECEC mode has one fewer function evaluation than PECECE mode. More generally, if the corrector is run k times, the method is in P(EC)k or P(EC)kE mode. If the corrector method is iterated until it converges, this could be called PE(CE)∞.
HAKMEM
HAKMEM, alternatively known as AI Memo 239, is a February 1972 "memo" (technical report) of the MIT AI Lab containing a wide variety of hacks, including useful and clever algorithms for mathematical computation, some number theory and schematic diagrams for hardware – in Guy L. Steele's words, "a bizarre and eclectic potpourri of technical trivia". Contributors included about two dozen members and associates of the AI Lab. The title of the report is short for "hacks memo", abbreviated to six upper case characters that would fit in a single PDP-10 machine word (using a six-bit character set). == History == HAKMEM is notable as an early compendium of algorithmic technique, particularly for its practical bent, and as an illustration of the wide-ranging interests of AI Lab people of the time, which included almost anything other than AI research. HAKMEM contains original work in some fields, notably continued fractions. == Introduction == Compiled with the hope that a record of the random things people do around here can save some duplication of effort -- except for fun. Here is some little known data which may be of interest to computer hackers. The items and examples are so sketchy that to decipher them may require more sincerity and curiosity than a non-hacker can muster. Doubtless, little of this is new, but nowadays it's hard to tell. So we must be content to give you an insight, or save you some cycles, and to welcome further contributions of items, new or used.
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.
The Citation Project
The Citation Project is a series of studies that measure and analyze first-year college writing students' source use and their ability to understand and implement sources within their own writing. The Citation Project reveals students' source-use habits and the issues that can be seen based on their lack of proper citation skills, such as the prevalence of plagiarism, institution policies, and the results of current writing pedagogy. The Citation Project's central findings were first presented at the Conference on College Composition and Communication in 2012. Although The Citation Project originally referred to this single 2012 study, the feedback received led to the conception of the Project as a broader initiative and as a place to gather and publish studies and data relating to student writing habits for the usage of other researches. == Method == The Citation Project's data comes from the work of 20 researchers analyzing 174 first-year composition students' research papers. The student papers studied originated from 16 institutions across the United States of America, including community colleges, public and private universities, denominational colleges, and Ivy Leagues. Researchers used bibliographic coding to aggregate data regarding the type, length, reading level, and usage of students' sources. == Findings == === Student source assessment and use === This study found that students were capable of identifying, locating, and accessing librarian-approved academic sources, most commonly accessing them with the internet. Despite students demonstrating their ability to find appropriate sources, they tend to exclusively cite the first few pages of their sources. Students' use and analysis of their citations are often limited, frequently resorting to patchwriting, directly restating their source's points, and omitting their own interpretations of their reference's ideas. The Citation Project also highlights students' struggle to accurately determine, address, and value their sources' bias, authority, and credibility. According to the Project's researchers' analysis, these habits demonstrate that first-year college writing students minimally engage with their sources and the academic conversations between them. One researcher from the Citation Project, Rebecca Moore Howard, believes these findings do not point towards students being lazy, but is rather a result of a writing pedagogy that prioritizes efficient, product-focused writing. Another interpretation offered by Sandra Jamieson, another researcher from the Citation Project explains their findings as a result of a lack of adherence to Information Learning (IL) Standards. === Pedagogy === A significant focus of The Citation Project is the development of pedagogical practices intended to equip students with writing and research techniques that will set them up for future success. Writers associated with The Citation Project, such as Tricia Serviss, believe that the practices of teachers surrounding academic integrity and writing practices are what form the foundation of how students think about writing and how to engage with assignments throughout their academic career. They also stress the importance of teaching students to effectively engage with sources rather than simply how to correctly cite them. The Citation Project asserts that endowing students with the ability to read, understand, and synthesize a variety of sources in their writing is a skill that will benefit them throughout their academic careers, and that the surface level typographical focus that many writing programs utilize is inadequate. == Plagiarism == One of the areas that The Citation Project also looks at is how students commit plagiarism throughout their writing. Plagiarism tends to be a checkpoint that gives instructors a sense where students' citation skills stand. Findings from The Citation Project reveal that the most common type of plagiarism is patchwriting which is the act of using the same sentence with only changing a couple of words. These types of issues can be seen as a learning curve due to lack of proper training. Student's that commit plagiarism are often unaware. === Policies === Another issue found is that academic plagiarism policies may not benefit a student's growth but may instead obstruct it. Policies against plagiarism tend to be harsh on the student that committed of offense. Even though student plagiarism is often unintentional academic institutions see this behavior as intentional. Student may then face harsh consequences as a result from their lack of citation knowledge. Additionally, higher level institutions assume that new students already have the skill set to avoid plagiarism which may be an unrealistic expectation. == Legacy == === Inspired studies === ==== Parrott and Napier ==== In one study, "Critical Reading and Student Self-Selected Texts: Results of a Collaborative, Explicit Curricular Approach," Jill Parrot and Trenia Napier quoted the Citation Project's findings as evidence that current collegiate writing curriculums are an ineffective means of teaching students how to properly write academic research papers. The researchers accredited current writing pedagogy's lack of emphasis on teaching critical reading skills. Parrott and Napier tested their thesis by seeing if students would produce more academic writing if they partook in a writing course that taught critical reading. Their results mostly went against this hypothesis, finding students who received additional critical reading training only significantly improved in how they integrated their sources. ==== Kocatepe ==== In May Mehtap Kocatep's study, "Reconceptualising the notion of finding information: How undergraduate students construct information as they read-to-write in an academic writing class," Kocatep expresses that she believes current conversations around writing pedagogy, including the Citation Project, operate with the underlying misconception that information is an easily discoverable static entity and its retrieval is an objective, unbiased decision. Kocatepe instead offers the analysis of what students view as valuable information and if it is worth using is influenced by the socially constructed meanings available to writers at the moment. To further examine students' source engagement, Kocatepe did a study on how female university students from the United Arab Emirates find, retrieve, use, and value sources. Kocatepe's results mainly noted students' almost exclusive reliance on using Google to find sources, as well as how students' navigated mainly English-speaking academic conversations as non-native English speakers. ==== Dahlen, Nordstrom-Sanchez, and Graff ==== Dahlen, Nordstrom-Sanchez, and Graff built their study off The Citation Project research in order to explore the attitudes and practices of students in an undergraduate writing course. As the researchers acknowledge, data collected by the Citation Project was the subject of the bulk of their analysis. This study sought to examine undergraduate writing practices tied to source-usage and elucidate any relevant trends. Dahlen, Nordstrom-Sanchez and Graff found that undergraduate writing students were not engaging with outside sources properly. Key issues discussed include lack of engagement with broad source ideas (in favor of picking out quotes), lack of paraphrasing, and inability to link information between multiple sources. ==== Davis ==== Phillip M. Davis based much of the analysis in his study on data gathered by the Citation Project. This study aimed to examine the particular effects web-based research and study had on undergraduate's papers and the replicability of their bibliographies. Davis sought to see how the shift from physical in-person library based research to online, often at-home research changed the function and usability of the bibliography as a form of documenting source usage in a given work. The primary method of analysis involved examining students' bibliographies to see where they were finding information online and how these sources were accessed. A main issue Davis found was "persistency" of URLs used for online citations. He found that only 18% of URL-based citations continued to function (the others either no longer pointing to the correct document or ceasing to exist altogether) within 3 years of their usage by students, and more than half of claimed online citations could not be found in any form. He suggests that this result brings up questions about how web-based citations should be dealt with in a university context.