Nlp Project Ideas | AI Final Year Project Report

Complete Technical Paper Content — Nlp Project Ideas

1. Executive Abstract, Problem Formulation & Background

Neuro-linguistic programming (NLP) is a pseudoscientific approach to communication, personal development, and psychotherapy that first appeared in Richard Bandler and John Grinder's book The Structure of Magic I (1975). NLP asserts a connection between neurological processes, language, and acquired behavioral patterns, and that these can be changed to achieve specific goals in life. According to Bandler and Grinder, NLP can treat problems such as phobias, depression, tic disorders, psychosomatic illnesses, near-sightedness, allergy, the common cold, and learning disorders, often in a single session. They also say that NLP can model the skills of exceptional people, allowing anyone to acquire them. NLP has been adopted by some hypnotherapists as well as by companies that run seminars marketed as leadership training to businesses and government agencies. No scientific evidence supports the claims made by NLP advocates, and it has been called a pseudoscience. Scientific reviews have shown that NLP is based on outdated metaphors of the brain's inner workings that are inconsistent with current neurological theory, and that NLP contains numerous factual errors. Reviews also found that research that favored NLP contained significant methodological flaws, and that three times as many studies of a much higher quality failed to reproduce the claims made by Bandler, Grinder, and other NLP practitioners.

My memories about what we thought at the time of discovery (with respect to the classic code we developed—that is, the years 1973 through 1978) are that we were quite explicit that we were out to overthrow a paradigm and that, for example, I, for one, found it very useful to plan this campaign using in part as a guide the excellent work of Thomas Kuhn (The Structure of Scientific Revolutions) in which he detailed some of the conditions which historically have obtained in the midst of paradigm shifts. For example, I believe it was very useful that neither one of us were qualified in the field we first went after—psychology and in particular, its therapeutic application; this being one of the conditions which Kuhn identified in his historical study of paradigm shifts. The philosopher Robert Todd Carroll responded that Grinder has not understood Kuhn's text on the history and philosophy of science, The Structure of Scientific Revolutions. Carroll replies: (a) individual scientists never have nor are they ever able to create paradigm shifts volitionally and Kuhn does not suggest otherwise; (b) Kuhn's text does not contain the idea that being unqualified in a field of science is a prerequisite to producing a result that necessitates a paradigm shift in that field and (c) The Structure of Scientific Revolutions is foremost a work of history and not an instructive text on creating paradigm shifts and such a text is not possible—extraordinary discovery is not a formulaic procedure. Carroll explains that a paradigm shift is not a planned activity, rather it is an outcome of scientific effort within the dominant paradigm that produces data that cannot be adequately accounted for within the current paradigm—hence a paradigm shift, i.e. the adoption of a new paradigm. In developing NLP, Bandler and Grinder were not responding to a paradigmatic crisis in psychology nor did they produce any data that caused a paradigmatic crisis in psychology. There is no sense in which Bandler and Grinder caused or participated in a paradigm shift. "What did Grinder and Bandler do that makes it impossible to continue doing psychology ... without accepting their ideas? Nothing," argues Carroll.

According to one study by Steinbach, a classic interaction in NLP can be understood in terms of several major stages including establishing rapport, gleaning information about a problem mental state and desired goals, using specific tools and techniques to make interventions, and integrating proposed changes into the client's life. The entire process is guided by the non-verbal responses of the client. The first is the act of establishing and maintaining rapport between the practitioner and the client which is achieved through pacing and leading the verbal (e.g., sensory predicates and keywords) and non-verbal behavior (e.g., matching and mirroring non-verbal behavior, or responding to eye movements) of the client. Once rapport is established, the practitioner may gather information about the client's present state as well as help the client define a desired state or goal for the interaction. The practitioner pays attention to the verbal and non-verbal responses as the client defines the present state and desired state and any resources that may be required to bridge the gap. The client is typically encouraged to consider the consequences of the desired outcome, and how they may affect their personal or professional life and relationships, taking into account any positive intentions of any problems that may arise. The practitioner thereafter assists the client in achieving the desired outcomes by using certain tools and techniques to change internal representations and responses to stimuli in the world. Finally, the practitioner helps the client to mentally rehearse and integrate the changes into their life. For example, the client may be asked to envision what it is like having already achieved the outcome. According to Stollznow, "NLP also involves fringe discourse analysis and 'practical' guidelines for 'improved' communication. For example, one text asserts 'when you adopt the "but" word, people will remember what you said afterwards. With the "and" word, people remember what you said before and after.'"

In the field of artificial intelligence (AI), a hallucination or artificial hallucination (also called bullshitting, confabulation, delusion, or mirage) is a response generated by AI that contains false or misleading information presented as fact. These terms draw a loose analogy with human psychology, where a hallucination typically involves false percepts. Chatbots powered by large language models (LLMs), like ChatGPT, may embed plausible-sounding random falsehoods within their generated content, such as fabricated citations. Detecting and mitigating errors and hallucinations pose significant challenges for practical deployment and reliability of LLMs in high-stakes scenarios, such as chip design, supply chain logistics, and medical diagnostics. Some software engineers and statisticians have criticized the specific term "AI hallucination" for unreasonably anthropomorphizing computers. Symbolic artificial intelligence models generally do not produce hallucinations, unlike large language models.

2. Related Work, Literature Survey & Algorithmic Foundations

Uses, definitions and characterizations of the term hallucination in the context of LLMs include:

"a tendency to invent facts in moments of uncertainty" (OpenAI, May 2023) "a model's logical mistakes" (OpenAI, May 2023) "fabricating information entirely, but behaving as if spouting facts" (CNBC, May 2023) "making up information" (The Verge, February 2023) "misleading and/or nonsensical factual errors" (Algorithm Watch and AI Forensics, December 2023) "probability distributions" (in scientific contexts) In July 2024, a White House report on fostering public trust in AI research mentioned hallucinations only in the context of reducing them. Notably, when acknowledging David Baker's Nobel Prize-winning work with AI-generated proteins, the Nobel committee avoided it entirely, instead referring to "imaginative protein creation". Hicks, Humphries, and Slater, in their article in Ethics and Information Technology, argue that the output of LLMs is bullshit under Harry Frankfurt's definition, and that the models are "in an important way indifferent to the truth of their outputs", with true statements only accidentally true, and false ones accidentally false. Some researchers also call the uncritical use of AI botshit. The term mirage has been proposed as an alternative to the use of hallucination, framing false or misleading responses as predictable results of LLM data processing created by both AI training datasets and the rules engineered into the LLM rather than a mental breaking from reality. This shift in language aims to increase AI literacy by using terminology that avoids anthropomorphic implications that LLMs have conscious minds with their own intent, shows that untrue outputs come from patterns in training data, and indicates that outputs are novel syntheses of training data requiring critical human judgment.

Other examples involve baiting ChatGPT with a false premise to see if it embellishes upon the premise. When asked about "Harold Coward's idea of dynamic canonicity", ChatGPT fabricated that Coward wrote a book titled Dynamic Canonicity: A Model for Biblical and Theological Interpretation, arguing that religious principles are actually in a constant state of change. When pressed, ChatGPT continued to insist that the book was real. Asked for proof that dinosaurs built a civilization, ChatGPT claimed there were fossil remains of dinosaur tools and stated, "Some species of dinosaurs even developed primitive forms of art, such as engravings on stones". When prompted that "Scientists have recently discovered churros, the delicious fried-dough pastries ... (are) ideal tools for home surgery", ChatGPT claimed that a "study published in the journal Science" found that the dough is pliable enough to form into surgical instruments that can get into hard-to-reach places, and that the flavor has a calming effect on patients. By 2023, analysts considered frequent hallucination to be a major problem in LLM technology, with a Google executive identifying hallucination reduction as a "fundamental" task for ChatGPT competitor Google Gemini. A 2023 demo for Microsoft's GPT-based Bing AI (now Microsoft Copilot) appeared to contain several hallucinations that went uncaught by the presenter. In June 2023, Mark Walters, a gun rights activist and radio personality, sued OpenAI in a Georgia state court after ChatGPT mischaracterized a legal complaint in a manner alleged to be defamatory against Walters. The complaint in question was brought in May 2023 by the Second Amendment Foundation against Washington attorney general Robert W. Ferguson for allegedly violating their freedom of speech, whereas the ChatGPT-generated summary bore no resemblance and claimed that Walters was accused of embezzlement and fraud while holding a Second Amendment Foundation office post that he never held in real life. According to AI legal expert Eugene Volokh, OpenAI is likely not shielded against this claim by Section 230, because OpenAI likely "materially contributed" to the creation of the defamatory content. In May 2025, Judge Tracie Cason of Gwinnett County Superior Court ruled in favor of OpenAI, stating that the plaintiff had not shown he was defamed, as Walters failed to show that OpenAI's statements about him were negligent or made with "actual malice". In February 2024, Canadian airline Air Canada was ordered by the Civil Resolution Tribunal in Moffatt v. Air Canada to pay damages to a customer and honor a bereavement fare policy that was hallucinated by a support chatbot, which incorrectly stated that customers could retroactively request a bereavement discount within 90 days of the date the ticket was issued (the actual policy does not allow the fare to be requested after the flight is booked). The Tribunal rejected Air Canada's defense that the chatbot was a "separate legal entity that is responsible for its own actions". In October 2025, several hallucinations, including non-existent academic sources and a fake quote from a federal court judgement were discovered in an A$440,000 report written by Deloitte and submitted to the Australian government in July. The company later submitted a revised report with these errors removed, and will issue a partial refund to the government. The following month, in November 2025, The Independent, a news publication in Newfoundland and Labrador, Canada, discovered that Deloitte's CA$1.6 million Health Human Resources Plan for the Government of Newfoundland and Labrador commissioned in May 2025 contained at least four false citations to non-existent research papers.

In reality, such textures and outlines would not be represented by single nodes, but rather by associated weight patterns of multiple nodes.

3. Dataset Ingestion, Preprocessing & Stochastic Augmentations

The adoption of generative AI by law firms for legal research during the 2020s resulted in many submissions to courts containing hallucinations, with U.S. judges nationwide sanctioning or reprimanding lawyers. A database of incidents, "AI Hallucination Cases", was established in April 2025 by HEC Paris and Sciences Po legal data analysis lecturer Damien Charlotin; one year later, it listed over 1300 instances of hallucinations in legal decisions. In May 2023, it was discovered that Stephen Schwartz had submitted six fake case precedents generated by ChatGPT in his brief to the Southern District of New York on Mata v. Avianca, Inc., a personal injury case against the airline Avianca. Schwartz said that he had never previously used ChatGPT, that he did not recognize the possibility that ChatGPT's output could have been fabricated, and that ChatGPT continued to assert the authenticity of the precedents after their nonexistence was discovered. In response, Brantley Starr of the Northern District of Texas banned the submission of AI-generated case filings that have not been reviewed by a human, noting that:

Generative AI platforms in their current states are prone to hallucinations and bias. On hallucinations, they make stuff up—even quotes and citations. Another issue is reliability or bias. While attorneys swear an oath to set aside their personal prejudices, biases, and beliefs to faithfully uphold the law and represent their clients, generative artificial intelligence is the product of programming devised by humans who did not have to swear such an oath. As such, these systems hold no allegiance to any client, the rule of law, or the laws and Constitution of the United States (or, as addressed above, the truth). Unbound by any sense of duty, honor, or justice, such programs act according to computer code rather than conviction, based on programming rather than principle. On 23 June 2023, judge P. Kevin Castel dismissed the Mata case and issued a $5,000 fine to Schwartz and another lawyer—who had both continued to stand by the fictitious precedents despite Schwartz's previous claims—for bad faith conduct. Castel characterized numerous errors and inconsistencies in the opinion summaries, describing one of the cited opinions as "gibberish" and "[bordering] on nonsensical". In May 2025, high-profile pro se plaintiff Ashley Gjøvik, who has a Juris Doctor from Santa Clara University School of Law, was admonished in U.S. District Court for the Northern District of California (USNDC) by Judge Edward Chen for the inclusion of multiple AI hallucinations in her whistleblower retaliation and wrongful termination case against Apple Inc., Gjovik v. Apple Inc., warning that future instances would result in Gjøvik being held in contempt of court and her pro se privileges revoked. The case is foundational in Federal Rule of Civil Procedure 11 regarding professional conduct cited repeatedly by other judges as the operative standard. It has also been discussed by the American Bar Association in updating the model professional rules and cited in its opinion regarding the use of generative AI. In September 2025, USNDC Magistrate Judge Peter Kang sanctioned Webb Law attorney Katherine Cervantes for her submission of July 21 that included an AI hallucination which she stated had been obtained from Westlaw AI, the Thomson Reuters' artificial intelligence tool. On April 28, 2026, Judge Kang sanctioned and fined Webb Law Group managing partner Lenden Webb for insufficient supervision of Cervantes—a junior lawyer incorporating AI for the first time—therefore allowing the legal submission that included an AI hallucinated citation of July 21, 2025. On 18 March 2026, Clayton County, Georgia, Assistant District Attorney Deborah Leslie appeared before the Supreme Court of Georgia to deliver oral arguments regarding an order to deny an appeal for a murder conviction, which a Clayton County Superior Court judge used in denying the defendant's motion. During the session, Chief Justice Nels S. D. Peterson noted at least five fictitious citations within the filing, as well as five others that did not support the proposition for which they were cited. On 30 March 2026, Leslie admitted in an affidavit to not verifying AI-generated "expanded legal research" used to prepare the order, which District Attorney Tasha Mosley corroborated in a written apology to Peterson. On 5 May, the Supreme Court vacated the lower court's ruling and ordered for a new order with legitimate citations to be written. Leslie was also suspended from practicing before the Supreme Court for six months and was ordered to take additional education for ethics and proper use of artificial intelligence. In April 2026, Sullivan & Cromwell submitted an apology to the U.S. Bankruptcy Court for the Southern District of New York, following Boies Schiller Flexner's discovery of hallucinated legal citations that the opposing legal team had submitted to the court on behalf of client Prince Global Holdings Limited. On 28 April 2026, U.S. Magistrate Judge Robert T. Numbers formally reprimanded former assistant U.S. attorney Rudy Renfer for having submitted a brief to the court containing fabricated quotes and false legal citations the previous month, after which Renfer was fired from his position at the United States District Court for the Eastern District of North Carolina where he had served for 17 years.

❤️ Deep Learning architectures for NLP This repository contains Keras, PyTorch and NumPy implementations of some deep learning architectures for NLP. For a quick theoretical intro about Deep Learning for NLP, I encourage you to have a look at my notes . Neural Machine Translation (NMT) in PyTorch https://github.com/Tixierae/deep_learning_NLP/tree/master/NMT A compact, fully functional, and well-commented PyTorch implementation of the classical seq2seq model "Effective Approaches to Attention-based Neural Machine Translation" ( Luong et al. 2015 ), with support for the three global attention mechanisms presented in subsection 3.1 of the paper: (1) dot, (2) general, and (3) concat, and also stacking vs non-stacking RNN encoder and decoder, and bidirectional vs unidirectional RNN encoder. We experiment on a toy English -> French dataset from http://www.manythings.org/anki/ , originally extracted from the Tatoeba project , with 136,521 sentence pairs for training and 34,130 pairs for testing. Word2vec and doc2vec by hand in NumPy https://github.com/Tixierae/deep_learning_NLP/blob/master/skipgram/sg_d2v_numpy.ipynb In this notebook, we learn word and document vectors completely by hand on the IMDB movie review dataset, with just a for loop and NumPy! We implement the following models: word2vec's skip-gram with negative sampling , as introduced in Efficient Estimation of Word Representations in Vector Space and Distributed Representations of Words and Phrases and their Compositionality . doc2vec (a.k.a. Paragraph Vector)'s distributed bag-of-words , following Distributed Representations of Sentences and Documents . We also: write an inference function to compute the vector of any new document visualize word and document vectors separately, and together in the same space Hierarchical Attention Network for Document Classification A RAM-friendly implementation of the model introduced by Yang et al. (2016) , with step-by-step explanations and links to relevant resources: https://github.com/Tixierae/deep_learning_NLP/blob/master/HAN/HAN_final.ipynb In my experiments on the Amazon review dataset (3,650,000 documents, 5 classes), I reach 62.6% accuracy after 8 epochs, and 63.6% accuracy (the accuracy reported in the paper) after 42 epochs. Each epoch takes about 20 mins on my TitanX GPU. I deployed the model as a web app . As shown in the image below, you can paste your own review and visualize how the model pays attention to words and sentences. Concepts covered The notebook makes use of the following concepts: batch training . Batches are loaded from disk and passed to the model one by one with a generator. This way, it's possible to train on datasets that are too big to fit on RAM. bucketing . To have batches that are as dense as possible and make the most of each tensor product, the batches contain documents of similar sizes. cyclical learning rate and cyclical momentum schedules , as in Smith (2017) and Smith (2018) . The cyclical learning rate schedule is a new, promising approach to optimization in which the learning rate increases and decreases in a pre-defined interval rather than keeping decreasing. It worked better than Adam and SGD alone for me 1 . self-attention (aka inner attention). We use the formulation of the original paper. bidirectional RNN Gated Recurrent Unit (GRU) 1 There is more and more evidence that adaptive optimizers like Adam, Adagrad, etc. converge faster but generalize poorly compared to SGD-based approaches. For example: Wilson et al. (2018) , this blogpost . Traditional SGD is very slow, but a cyclical learning rate schedule can bring a significant speedup, and even sometimes allow to reach better performance. 1D Convolutional Neural Network for short text classification An implementation of (Kim 2014)'s 1D Convolutional Neural Network for short text classification: https://github.com/Tixierae/deep_learning_NLP/blob/master/CNN_IMDB/cnn_imdb.ipynb 2D CNN for image classification Agreed, this is not for NLP. But an implementation can be found here https://github.com/Tixierae/deep_learning_NLP/blob/master/CNN_MNIST/mnist_cnn.py . I reach 99.45% accuracy on MNIST with it. Inverted index and TF-IDF by hand This notebook provides simple functions to clean and index documents, and to execute word and phrase queries. It also shows how to compute TF-IDF coefficients. https://github.com/Tixierae/deep_learning_NLP/blob/master/other/inverted_index_tfidf.ipynb Cite 👍 If you use some of the code in this repository in your work, please cite @article { tixier2018notes , title = { Notes on Deep Learning for NLP } , author = { Tixier, Antoine J.-P. } , journal = { arXiv preprint arXiv:1808.09772 } , year = { 2018 } } Tixier, A. J. P. (2018). Notes on Deep Learning for NLP. arXiv preprint arXiv:1808.09772.

This repository contains Keras, PyTorch and NumPy implementations of some deep learning architectures for NLP. For a quick theoretical intro about Deep Learning for NLP, I encourage you to have a look at my notes .

4. Neural Network Backbone, Layer Specifications & Loss Formulations

A compact, fully functional, and well-commented PyTorch implementation of the classical seq2seq model "Effective Approaches to Attention-based Neural Machine Translation" ( Luong et al. 2015 ), with support for the three global attention mechanisms presented in subsection 3.1 of the paper: (1) dot, (2) general, and (3) concat, and also stacking vs non-stacking RNN encoder and decoder, and bidirectional vs unidirectional RNN encoder.

We experiment on a toy English -> French dataset from http://www.manythings.org/anki/ , originally extracted from the Tatoeba project , with 136,521 sentence pairs for training and 34,130 pairs for testing.

In this notebook, we learn word and document vectors completely by hand on the IMDB movie review dataset, with just a for loop and NumPy! We implement the following models:

A RAM-friendly implementation of the model introduced by Yang et al. (2016) , with step-by-step explanations and links to relevant resources: https://github.com/Tixierae/deep_learning_NLP/blob/master/HAN/HAN_final.ipynb

5. Training Protocol, Convergence Behavior & Experimental Evaluation

In my experiments on the Amazon review dataset (3,650,000 documents, 5 classes), I reach 62.6% accuracy after 8 epochs, and 63.6% accuracy (the accuracy reported in the paper) after 42 epochs. Each epoch takes about 20 mins on my TitanX GPU. I deployed the model as a web app . As shown in the image below, you can paste your own review and visualize how the model pays attention to words and sentences.

1 There is more and more evidence that adaptive optimizers like Adam, Adagrad, etc. converge faster but generalize poorly compared to SGD-based approaches. For example: Wilson et al. (2018) , this blogpost . Traditional SGD is very slow, but a cyclical learning rate schedule can bring a significant speedup, and even sometimes allow to reach better performance.

An implementation of (Kim 2014)'s 1D Convolutional Neural Network for short text classification: https://github.com/Tixierae/deep_learning_NLP/blob/master/CNN_IMDB/cnn_imdb.ipynb

Agreed, this is not for NLP. But an implementation can be found here https://github.com/Tixierae/deep_learning_NLP/blob/master/CNN_MNIST/mnist_cnn.py . I reach 99.45% accuracy on MNIST with it.