Recent methods have demonstrated that Large Language Models (LLMs) can solve reasoning tasks better when they are encouraged to solve subtasks of the main task first. In this paper we devise a similar strategy that breaks down reasoning tasks into a problem decomposition phase and a problem solving phase and show that the strategy is able to outperform a single stage solution. Further, we hypothesize that the decomposition should be easier to distill into a smaller model compared to the problem solving because the latter requires large amounts of domain knowledge while the former only requires learning general problem solving strategies. We propose methods to distill these two capabilities and evaluate their impact on reasoning outcomes and inference cost. We find that we can distill the problem decomposition phase and at the same time achieve good generalization across tasks, datasets, and models. However, it is harder to distill the problem solving capability without losing performance and the resulting distilled model struggles with generalization. These results indicate that by using smaller, distilled problem decomposition models in combination with problem solving LLMs we can achieve reasoning with cost-efficient inference and local adaptation.

Textual backdoor attacks, characterized by subtle manipulations of input triggers and training dataset labels, pose significant threats to security-sensitive applications. The rise of advanced generative models, such as GPT-4, with their capacity for human-like rewriting, makes these attacks increasingly challenging to detect. In this study, we conduct an in-depth examination of black-box generative models as tools for backdoor attacks, thereby emphasizing the need for effective defense strategies. We propose BGMAttack, a novel framework that harnesses advanced generative models to execute stealthier backdoor attacks on text classifiers. Unlike prior approaches constrained by subpar generation quality, BGMAttack renders backdoor triggers more elusive to human cognition and advanced machine detection. A rigorous evaluation of attack effectiveness over four sentiment classification tasks, complemented by four human cognition stealthiness tests, reveals BGMAttack’s superior performance, achieving a state-of-the-art attack success rate of 97.35% on average while maintaining superior stealth compared to conventional methods. The dataset and code are available: https://github.com/JiazhaoLi/BGMAttack.

Adverse drug events (ADEs) pose major public health risks, with traditional reporting systems often failing to capture them. Our proposed pipeline, called Deep-LLMADEminer, used natural language processing approaches to tackle this issue for #SMM4H 2024 shared task 1. Using annotated tweets, we built a three part pipeline: RoBERTa for classification, GPT-4-turbo for span extraction, and BioBERT for normalization. Our models achieved F1-scores of 0.838, 0.306, and 0.354, respectively, offering a novel system for Task 1 and similar pharmacovigilance tasks.

This paper summarizes our participation in the Shared Task 4 of #SMM4H 2024. Task 4 was a named entity recognition (NER) task identifying clinical and social impacts of non-medical substance use in English Reddit posts. We employed the Bidirectional Encoder Representations from Transformers (BERT) model to complete this task. Our team achieved an F1-score of 0.892 on a validation set and a relaxed F1-score of 0.191 on the test set.

The goal of Social Media Mining for Health (#SMM4H) 2024 Task 7 was to train a machine learning model that is able to distinguish between annotations made by humans and those made by a Large Language Model (LLM). The dataset consisted of tweets originating from #SMM4H 2023 Task 3, wherein the objective was to extract COVID-19 symptoms in Latin-American Spanish tweets. Due to the lack of additional annotated tweets for classification, we reframed the task using the available tweets and their corresponding human or machine annotator labels to explore differences between the two subsets of tweets. We conducted an exploratory data analysis and trained a BERT-based classifier to identify sampling biases between the two subsets. The exploratory data analysis found no significant differences between the samples and our best classifier achieved a precision of 0.52 and a recall of 0.51, indicating near-random performance. This confirms the lack of sampling biases between the two sets of tweets and is thus a valid dataset for a task designed to assess the authorship of annotations by humans versus machines.

Generalization refers to the ability of machine learning models to perform well on dataset distributions different from the one it was trained on. While several pre-existing works have characterized the generalizability of NLP models across different dimensions, such as domain shift, adversarial perturbations, or compositional variations, most studies were carried out in a stand-alone setting, emphasizing a single dimension of interest. We bridge this gap by systematically investigating the generalizability of pre-trained language models across different architectures, sizes, and training strategies, over multiple dimensions for the task of natural language inference and question answering. Our results indicate that model instances typically exhibit consistent generalization trends, i.e., they generalize equally well (or poorly) across most scenarios, and this ability is correlated with model architecture, base dataset performance, size, and training mechanism. We hope this research motivates further work in a) developing a multi-dimensional generalization benchmark for systematic evaluation and b) examining the reasons behind models’ generalization abilities. The code and data are available at https://github.com/sagnik/md-gen-nlp, and the trained models are released at https://huggingface.co/varun-v-rao.

As part of our participation in BioLaySumm 2023, we explored the use of large language models (LLMs) to automatically generate concise and readable summaries of biomedical research articles. We utilized pre-trained LLMs to fine-tune our summarization models on two provided datasets, and adapt them to the shared task within the constraints of training time and computational power. Our final models achieved very high relevance and factuality scores on the test set, and ranked among the top five models in the overall performance.

Textual backdoor attack, as a novel attack model, has been shown to be effective in adding a backdoor to the model during training. Defending against such backdoor attacks has become urgent and important. In this paper, we propose AttDef, an efficient attribution-based pipeline to defend against two insertion-based poisoning attacks, BadNL and InSent. Specifically, we regard the tokens with larger attribution scores as potential triggers since larger attribution words contribute more to the false prediction results and therefore are more likely to be poison triggers. Additionally, we further utilize an external pre-trained language model to distinguish whether input is poisoned or not. We show that our proposed method can generalize sufficiently well in two common attack scenarios (poisoning training data and testing data), which consistently improves previous methods. For instance, AttDef can successfully mitigate both attacks with an average accuracy of 79.97% (56.59% up) and 48.34% (3.99% up) under pre-training and post-training attack defense respectively, achieving the new state-of-the-art performance on prediction recovery over four benchmark datasets.

Prompt tuning is a new, efficient NLP transfer learning paradigm that adds a task-specific prompt in each input instance during the model training stage. It freezes the pre-trained language model and only optimizes a few task-specific prompts. In this paper, we propose a conditional prompt generation method to generate prompts for each input instance, referred to as the Instance-Dependent Prompt Generation (IDPG). Unlike traditional prompt tuning methods that use a fixed prompt, IDPG introduces a lightweight and trainable component to generate prompts based on each input sentence. Extensive experiments on ten natural language understanding (NLU) tasks show that the proposed strategy consistently outperforms various prompt tuning baselines and is on par with other efficient transfer learning methods such as Compacter while tuning far fewer model parameters.

The team from the University of Michigan participated in three tasks in the Social Media Mining for Health Applications (#SMM4H) 2020 shared tasks – on detecting mentions of adverse effects (Task 2), extracting and normalizing them (Task 3), and detecting mentions of medication abuse (Task 4). Our approaches relied on a combination of traditional machine learning and deep learning models. On Tasks 2 and 4, our submitted runs performed at or above the task average.

The language used by physicians and health professionals in prescription directions includes medical jargon and implicit directives and causes much confusion among patients. Human intervention to simplify the language at the pharmacies may introduce additional errors that can lead to potentially severe health outcomes. We propose a novel machine translation-based approach, PharmMT, to automatically and reliably simplify prescription directions into patient-friendly language, thereby significantly reducing pharmacist workload. We evaluate the proposed approach over a dataset consisting of over 530K prescriptions obtained from a large mail-order pharmacy. The end-to-end system achieves a BLEU score of 60.27 against the reference directions generated by pharmacists, a 39.6% relative improvement over the rule-based normalization. Pharmacists judged 94.3% of the simplified directions as usable as-is or with minimal changes. This work demonstrates the feasibility of a machine translation-based tool for simplifying prescription directions in real-life.

Identifying mentions of medical concepts in social media is challenging because of high variability in free text. In this paper, we propose a novel neural network architecture, the Collocated LSTM with Attentive Pooling and Aggregated representation (CLAPA), that integrates a bidirectional LSTM model with attention and pooling strategy and utilizes the collocation information from training data to improve the representation of medical concepts. The collocation and aggregation layers improve the model performance on the task of identifying mentions of adverse drug events (ADE) in tweets. Using the dataset made available as part of the workshop shared task, we show that careful selection of neighborhood contexts can help uncover useful local information and improve the overall medical concept representation.

We participated in Task 1 of the Social Media Mining for Health Applications (SMM4H) 2019 Shared Tasks on detecting mentions of adverse drug events (ADEs) in tweets. Our approach relied on a text processing pipeline for tweets, and training traditional machine learning and deep learning models. Our submitted runs performed above average for the task.

In this paper we introduce the problem of identifying usage expression sentences in a consumer product review. We create a human-annotated gold standard dataset of 565 reviews spanning five distinct product categories. Our dataset consists of more than 3,000 annotated sentences. We further introduce a classification system to label sentences according to whether or not they describe some “usage”. The system combines lexical, syntactic, and semantic features in a product-agnostic fashion to yield good classification performance. We show the effectiveness of our approach using importance ranking of features, error analysis, and cross-product classification experiments.