In this paper, we investigate the use of data obtained from prompting a large generative language model, ChatGPT, to generate synthetic training data with the aim of augmenting data in low resource scenarios. We show that with appropriate task-specific ChatGPT prompts, we outperform the most popular existing approaches for such data augmentation. Furthermore, we investigate methodologies for evaluating the similarity of the augmented data generated from ChatGPT with the aim of validating and assessing the quality of the data generated.
Large Language Models (LLMs) are constrained by their inability to process lengthy inputs, resulting in the loss of critical historical information. To address this limitation, in this paper, we propose the Self-Controlled Memory (SCM) framework to enhance the ability of LLMs to maintain long-term memory and recall relevant information. Our SCM framework comprises three key components: an LLM-based agent serving as the backbone of the framework, a memory stream storing agent memories, and a memory controller updating memories and determining when and how to utilize memories from memory stream. Additionally, the proposed SCM is able to process ultra-long texts without any modification or fine-tuning, which can integrate with any instruction following LLMs in a plug-and-play paradigm. Furthermore, we annotate a dataset to evaluate the effectiveness of SCM for handling lengthy inputs. The annotated dataset covers three tasks: long-term dialogues, book summarization, and meeting summarization. Experimental results demonstrate that our method achieves better retrieval recall and generates more informative responses compared to competitive baselines in long-term dialogues. (https://github.com/wbbeyourself/SCM4LLMs)
Current query expansion models use pseudo-relevance feedback to improve first-pass retrieval effectiveness; however, this fails when the initial results are not relevant. Instead of building a language model from retrieved results, we propose Generative Relevance Feedback (GRF) that builds probabilistic feedback models from long-form text generated from Large Language Models. We study the effective methods for generating text by varying the zero-shot generation subtasks: queries, entities, facts, news articles, documents, and essays. We evaluate GRF on document retrieval benchmarks covering a diverse set of queries and document collections, and the results show that GRF methods significantly outperform previous PRF methods. Specifically, we improve MAP between 5-19% and NDCG@10 17-24% compared to RM3 expansion, and achieve the best R@1k effectiveness on all datasets compared to state-of-the-art sparse, dense, and expansion models.
Human language is full of compositional syntactic structures, and although neural networks have contributed to groundbreaking improvements in computer systems that process language, widely-used neural network architectures still exhibit limitations in their ability to process syntax. To address this issue, prior work has proposed adding stack data structures to neural networks, drawing inspiration from theoretical connections between syntax and stacks. However, these methods employ deterministic stacks that are designed to track one parse at a time, whereas syntactic ambiguity, which requires a nondeterministic stack to parse, is extremely common in language. In this dissertation, we remedy this discrepancy by proposing a method of incorporating nondeterministic stacks into neural networks. We develop a differentiable data structure that efficiently simulates a nondeterministic pushdown automaton, representing an exponential number of computations with a dynamic programming algorithm. We incorporate this module into two predominant architectures: recurrent neural networks (RNNs) and transformers. We show that this raises their formal recognition power to arbitrary context-free languages, and also aids training, even on deterministic context-free languages. Empirically, neural networks with nondeterministic stacks learn context-free languages much more effectively than prior stack-augmented models, including a language with theoretically maximal parsing difficulty. We also show that an RNN augmented with a nondeterministic stack is capable of surprisingly powerful behavior, such as learning cross-serial dependencies, a well-known non-context-free pattern. We demonstrate improvements on natural language modeling and provide analysis on a syntactic generalization benchmark. This work represents an important step toward building systems that learn to use syntax in more human-like fashion.
This paper describes the system entered by the author to the SemEval-2023 Task 12: Sentiment analysis for African languages. The system focuses on the Kinyarwanda language and uses a language-specific model. Kinyarwanda morphology is modeled in a two tier transformer architecture and the transformer model is pre-trained on a large text corpus using multi-task masked morphology prediction. The model is deployed on an experimental platform that allows users to experiment with the pre-trained language model fine-tuning without the need to write machine learning code. Our final submission to the shared task achieves second ranking out of 34 teams in the competition, achieving 72.50% weighted F1 score. Our analysis of the evaluation results highlights challenges in achieving high accuracy on the task and identifies areas for improvement.
This report describes GMU's sentiment analysis system for the SemEval-2023 shared task AfriSenti-SemEval. We participated in all three sub-tasks: Monolingual, Multilingual, and Zero-Shot. Our approach uses models initialized with AfroXLMR-large, a pre-trained multilingual language model trained on African languages and fine-tuned correspondingly. We also introduce augmented training data along with original training data. Alongside finetuning, we perform phylogeny-based adapter tuning to create several models and ensemble the best models for the final submission. Our system achieves the best F1-score on track 5: Amharic, with 6.2 points higher F1-score than the second-best performing system on this track. Overall, our system ranks 5th among the 10 systems participating in all 15 tracks.
This paper presents a dynamic, real-time approach to detecting anomalous blockchain transactions. The proposed tool, BlockGPT, generates tracing representations of blockchain activity and trains from scratch a large language model to act as a real-time Intrusion Detection System. Unlike traditional methods, BlockGPT is designed to offer an unrestricted search space and does not rely on predefined rules or patterns, enabling it to detect a broader range of anomalies. We demonstrate the effectiveness of BlockGPT through its use as an anomaly detection tool for Ethereum transactions. In our experiments, it effectively identifies abnormal transactions among a dataset of 68M transactions and has a batched throughput of 2284 transactions per second on average. Our results show that, BlockGPT identifies abnormal transactions by ranking 49 out of 124 attacks among the top-3 most abnormal transactions interacting with their victim contracts. This work makes contributions to the field of blockchain transaction analysis by introducing a custom data encoding compatible with the transformer architecture, a domain-specific tokenization technique, and a tree encoding method specifically crafted for the Ethereum Virtual Machine (EVM) trace representation.
In most natural language inference problems, sentence representation is needed for semantic retrieval tasks. In recent years, pre-trained large language models have been quite effective for computing such representations. These models produce high-dimensional sentence embeddings. An evident performance gap between large and small models exists in practice. Hence, due to space and time hardware limitations, there is a need to attain comparable results when using the smaller model, which is usually a distilled version of the large language model. In this paper, we assess the model distillation of the sentence representation model Sentence-BERT by augmenting the pre-trained distilled model with a projection layer additionally learned on the Maximum Coding Rate Reduction (MCR2)objective, a novel approach developed for general-purpose manifold clustering. We demonstrate that the new language model with reduced complexity and sentence embedding size can achieve comparable results on semantic retrieval benchmarks.
Structured State Spaces for Sequences (S4) is a recently proposed sequence model with successful applications in various tasks, e.g. vision, language modeling, and audio. Thanks to its mathematical formulation, it compresses its input to a single hidden state, and is able to capture long range dependencies while avoiding the need for an attention mechanism. In this work, we apply S4 to Machine Translation (MT), and evaluate several encoder-decoder variants on WMT'14 and WMT'16. In contrast with the success in language modeling, we find that S4 lags behind the Transformer by approximately 4 BLEU points, and that it counter-intuitively struggles with long sentences. Finally, we show that this gap is caused by S4's inability to summarize the full source sentence in a single hidden state, and show that we can close the gap by introducing an attention mechanism.
The proliferation of Large Language Models (LLMs), such as ChatGPT, has raised concerns about their potential impact on academic integrity, prompting the need for LLM-resistant exam designs. This article investigates the performance of LLMs on exams and their implications for assessment, focusing on ChatGPT's abilities and limitations. We propose guidelines for creating LLM-resistant exams, including content moderation, deliberate inaccuracies, real-world scenarios beyond the model's knowledge base, effective distractor options, evaluating soft skills, and incorporating non-textual information. The article also highlights the significance of adapting assessments to modern tools and promoting essential skills development in students. By adopting these strategies, educators can maintain academic integrity while ensuring that assessments accurately reflect contemporary professional settings and address the challenges and opportunities posed by artificial intelligence in education.
Query-focused summarization (QFS) aims to provide a summary of a document that satisfies information need of a given query and is useful in various IR applications, such as abstractive snippet generation. Current QFS approaches typically involve injecting additional information, e.g. query-answer relevance or fine-grained token-level interaction between a query and document, into a finetuned large language model. However, these approaches often require extra parameters \& training, and generalize poorly to new dataset distributions. To mitigate this, we propose leveraging a recently developed constrained generation model Neurological Decoding (NLD) as an alternative to current QFS regimes which rely on additional sub-architectures and training. We first construct lexical constraints by identifying important tokens from the document using a lightweight gradient attribution model, then subsequently force the generated summary to satisfy these constraints by directly manipulating the final vocabulary likelihood. This lightweight approach requires no additional parameters or finetuning as it utilizes both an off-the-shelf neural retrieval model to construct the constraints and a standard generative language model to produce the QFS. We demonstrate the efficacy of this approach on two public QFS collections achieving near parity with the state-of-the-art model with substantially reduced complexity.
Domain knowledge refers to the in-depth understanding, expertise, and familiarity with a specific subject, industry, field, or area of special interest. The existing benchmarks are all lack of an overall design for domain knowledge evaluation. Holding the belief that the real ability of domain language understanding can only be fairly evaluated by an comprehensive and in-depth benchmark, we introduces the Domma, a Domain Mastery Benchmark. DomMa targets at testing Large Language Models (LLMs) on their domain knowledge understanding, it features extensive domain coverage, large data volume, and a continually updated data set based on Chinese 112 first-level subject classifications. DomMa consist of 100,000 questions in both Chinese and English sourced from graduate entrance examinations and undergraduate exams in Chinese college. We have also propose designs to make benchmark and evaluation process more suitable to LLMs.
This is a tutorial paper on Recurrent Neural Network (RNN), Long Short-Term Memory Network (LSTM), and their variants. We start with a dynamical system and backpropagation through time for RNN. Then, we discuss the problems of gradient vanishing and explosion in long-term dependencies. We explain close-to-identity weight matrix, long delays, leaky units, and echo state networks for solving this problem. Then, we introduce LSTM gates and cells, history and variants of LSTM, and Gated Recurrent Units (GRU). Finally, we introduce bidirectional RNN, bidirectional LSTM, and the Embeddings from Language Model (ELMo) network, for processing a sequence in both directions.
This paper highlights the importance of personalization in large language models and introduces the LaMP benchmark -- a novel benchmark for training and evaluating language models for producing personalized outputs. LaMP offers a comprehensive evaluation framework with diverse language tasks and multiple entries for each user profile. It consists of seven personalized tasks, spanning three text classification and four text generation tasks. We additionally propose two retrieval augmentation approaches that retrieve personal items from each user profile for personalizing language model outputs. To this aim, we study various retrieval models, including term matching, semantic matching, and time-aware methods. Extensive experiments on LaMP for zero-shot and fine-tuned language models demonstrate the efficacy of the proposed retrieval augmentation approach and highlight the impact of personalization in various natural language tasks.
Recent psycholinguistic studies have drawn conflicting conclusions about the relationship between the quality of a language model and the ability of its surprisal estimates to predict human reading times, which has been speculated to be due to the large gap in both the amount of training data and model capacity across studies. The current work aims to consolidate these findings by evaluating surprisal estimates from Transformer-based language model variants that vary systematically in the amount of training data and model capacity on their ability to predict human reading times. The results show that surprisal estimates from most variants with contemporary model capacities provide the best fit after seeing about two billion training tokens, after which they begin to diverge from humanlike expectations. Additionally, newly-trained smaller model variants reveal a 'tipping point' at convergence, after which the decrease in language model perplexity begins to result in poorer fits to human reading times. These results suggest that the massive amount of training data is mainly responsible for the poorer fit achieved by surprisal from larger pre-trained language models, and that a certain degree of model capacity is necessary for Transformer-based language models to capture humanlike expectations.
Legal case retrieval, which aims to find relevant cases for a query case, plays a core role in the intelligent legal system. Despite the success that pre-training has achieved in ad-hoc retrieval tasks, effective pre-training strategies for legal case retrieval remain to be explored. Compared with general documents, legal case documents are typically long text sequences with intrinsic logical structures. However, most existing language models have difficulty understanding the long-distance dependencies between different structures. Moreover, in contrast to the general retrieval, the relevance in the legal domain is sensitive to key legal elements. Even subtle differences in key legal elements can significantly affect the judgement of relevance. However, existing pre-trained language models designed for general purposes have not been equipped to handle legal elements. To address these issues, in this paper, we propose SAILER, a new Structure-Aware pre-traIned language model for LEgal case Retrieval. It is highlighted in the following three aspects: (1) SAILER fully utilizes the structural information contained in legal case documents and pays more attention to key legal elements, similar to how legal experts browse legal case documents. (2) SAILER employs an asymmetric encoder-decoder architecture to integrate several different pre-training objectives. In this way, rich semantic information across tasks is encoded into dense vectors. (3) SAILER has powerful discriminative ability, even without any legal annotation data. It can distinguish legal cases with different charges accurately. Extensive experiments over publicly available legal benchmarks demonstrate that our approach can significantly outperform previous state-of-the-art methods in legal case retrieval.
Language models have become very popular recently and many claims have been made about their abilities, including for commonsense reasoning. Given the increasingly better results of current language models on previous static benchmarks for commonsense reasoning, we explore an alternative dialectical evaluation. The goal of this kind of evaluation is not to obtain an aggregate performance value but to find failures and map the boundaries of the system. Dialoguing with the system gives the opportunity to check for consistency and get more reassurance of these boundaries beyond anecdotal evidence. In this paper we conduct some qualitative investigations of this kind of evaluation for the particular case of spatial reasoning (which is a fundamental aspect of commonsense reasoning). We conclude with some suggestions for future work both to improve the capabilities of language models and to systematise this kind of dialectical evaluation.
The release of ChatGPT, Gemini, and other large language model has drawn huge interests on foundations models. There is a broad consensus that foundations models will be the fundamental building blocks for future AI systems. However, there is a lack of systematic guidance on the architecture design. Particularly, the the rapidly growing capabilities of foundations models can eventually absorb other components of AI systems, posing challenges of moving boundary and interface evolution in architecture design. Furthermore, incorporating foundations models into AI systems raises significant concerns about responsible and safe AI due to their opaque nature and rapidly advancing intelligence. To address these challenges, the paper first presents an architecture evolution of AI systems in the era of foundation models, transitioning from "foundation-model-as-a-connector" to "foundation-model-as-a-monolithic architecture". The paper then identifies key design decisions and proposes a pattern-oriented reference architecture for designing responsible foundation-model-based systems. The patterns can enable the potential of foundation models while ensuring associated risks.
In recent years, Large Language Models such as GPT-3 showed remarkable capabilities in performing NLP tasks in the zero and few shot settings. On the other hand, the experiments highlighted the difficulty of GPT-3 in carrying out tasks that require a certain degree of reasoning, such as arithmetic operations. In this paper we evaluate the ability of Transformer Language Models to perform arithmetic operations following a pipeline that, before performing computations, decomposes numbers in units, tens, and so on. We denote the models fine-tuned with this pipeline with the name Calculon and we test them in the task of performing additions, subtractions and multiplications on the same test sets of GPT-3. Results show an increase of accuracy of 63% in the five-digit addition task. Moreover, we demonstrate the importance of the decomposition pipeline introduced, since fine-tuning the same Language Model without decomposing numbers results in 0% accuracy in the five-digit addition task.
Pre-trained contextual language models are ubiquitously employed for language understanding tasks, but are unsuitable for resource-constrained systems. Noncontextual word embeddings are an efficient alternative in these settings. Such methods typically use one vector to encode multiple different meanings of a word, and incur errors due to polysemy. This paper proposes a two-stage method to distill multiple word senses from a pre-trained language model (BERT) by using attention over the senses of a word in a context and transferring this sense information to fit multi-sense embeddings in a skip-gram-like framework. We demonstrate an effective approach to training the sense disambiguation mechanism in our model with a distribution over word senses extracted from the output layer embeddings of BERT. Experiments on the contextual word similarity and sense induction tasks show that this method is superior to or competitive with state-of-the-art multi-sense embeddings on multiple benchmark data sets, and experiments with an embedding-based topic model (ETM) demonstrates the benefits of using this multi-sense embedding in a downstream application.