Unpaired text and audio injection have emerged as dominant methods for improving ASR performance in the absence of a large labeled corpus. However, little guidance exists on deploying these methods to improve production ASR systems that are trained on very large supervised corpora and with realistic requirements like a constrained model size and CPU budget, streaming capability, and a rich lattice for rescoring and for downstream NLU tasks. In this work, we compare three state-of-the-art semi-supervised methods encompassing both unpaired text and audio as well as several of their combinations in a controlled setting using joint training. We find that in our setting these methods offer many improvements beyond raw WER, including substantial gains in tail-word WER, decoder computation during inference, and lattice density.
The increasing popularity of large language models (LLMs) such as ChatGPT has led to growing concerns about their safety, security risks, and ethical implications. This paper aims to provide an overview of the different types of security risks associated with ChatGPT, including malicious text and code generation, private data disclosure, fraudulent services, information gathering, and producing unethical content. We present an empirical study examining the effectiveness of ChatGPT's content filters and explore potential ways to bypass these safeguards, demonstrating the ethical implications and security risks that persist in LLMs even when protections are in place. Based on a qualitative analysis of the security implications, we discuss potential strategies to mitigate these risks and inform researchers, policymakers, and industry professionals about the complex security challenges posed by LLMs like ChatGPT. This study contributes to the ongoing discussion on the ethical and security implications of LLMs, underscoring the need for continued research in this area.
Recent years witnessed an increase in the amount of research on the task of Question Difficulty Estimation from Text QDET with Natural Language Processing (NLP) techniques, with the goal of targeting the limitations of traditional approaches to question calibration. However, almost the entirety of previous research focused on single silos, without performing quantitative comparisons between different models or across datasets from different educational domains. In this work, we aim at filling this gap, by quantitatively analyzing several approaches proposed in previous research, and comparing their performance on three publicly available real world datasets containing questions of different types from different educational domains. Specifically, we consider reading comprehension Multiple Choice Questions (MCQs), science MCQs, and math questions. We find that Transformer based models are the best performing across different educational domains, with DistilBERT performing almost as well as BERT, and that they outperform other approaches even on smaller datasets. As for the other models, the hybrid ones often outperform the ones based on a single type of features, the ones based on linguistic features perform well on reading comprehension questions, while frequency based features (TF-IDF) and word embeddings (word2vec) perform better in domain knowledge assessment.
Large language models (LLMs) are becoming attractive as few-shot reasoners to solve NL-related tasks. However, there is still much to be learned about how well LLMs understand structured data, such as tables. While it is true that tables can be used as inputs to LLMs with serialization, there lack comprehensive studies examining whether LLMs can truly comprehend such data. In this paper we try to understand this by designing a benchmark to evaluate structural understanding capabilities (SUC) of LLMs. The benchmark we create includes seven tasks, each with their own unique challenges, e.g,, cell lookup, row retrieval and size detection. We run a series of evaluations on GPT-3 family models (e.g., text-davinci-003). We discover that the performance varied depending on a number of input choices, including table input format, content order, role prompting and partition marks. Drawing from the insights gained through the benchmark evaluations, we then propose self-augmentation for effective structural prompting, e.g., critical value / range identification using LLMs' internal knowledge. When combined with carefully chosen input choices, these structural prompting methods lead to promising improvements in LLM performance on a variety of tabular tasks, e.g., TabFact($\uparrow2.31\%$), HybridQA($\uparrow2.13\%$), SQA($\uparrow2.72\%$), Feverous($\uparrow0.84\%$), and ToTTo($\uparrow5.68\%$). We believe our benchmark and proposed prompting methods can serve as a simple yet generic selection for future research. The code and data are released in https://anonymous.4open.science/r/StructuredLLM-76F3.
While large language models (LMs) have shown remarkable capabilities across numerous tasks, they often struggle with simple reasoning and planning in physical environments, such as understanding object permanence or planning household activities. The limitation arises from the fact that LMs are trained only on written text and miss essential embodied knowledge and skills. In this paper, we propose a new paradigm of enhancing LMs by finetuning them with world models, to gain diverse embodied knowledge while retaining their general language capabilities. Our approach deploys an embodied agent in a world model, particularly a simulator of the physical world (VirtualHome), and acquires a diverse set of embodied experiences through both goal-oriented planning and random exploration. These experiences are then used to finetune LMs to teach diverse abilities of reasoning and acting in the physical world, e.g., planning and completing goals, object permanence and tracking, etc. Moreover, it is desirable to preserve the generality of LMs during finetuning, which facilitates generalizing the embodied knowledge across tasks rather than being tied to specific simulations. We thus further introduce the classical elastic weight consolidation (EWC) for selective weight updates, combined with low-rank adapters (LoRA) for training efficiency. Extensive experiments show our approach substantially improves base LMs on 18 downstream tasks by 64.28% on average. In particular, the small LMs (1.3B and 6B) enhanced by our approach match or even outperform much larger LMs (e.g., ChatGPT).
Handling large amounts of data has become a key for developing automated driving systems. Especially for developing highly automated driving functions, working with images has become increasingly challenging due to the sheer size of the required data. Such data has to satisfy different requirements to be usable in machine learning-based approaches. Thus, engineers need to fully understand their large image data sets for the development and test of machine learning algorithms. However, current approaches lack automatability, are not generic and are limited in their expressiveness. Hence, this paper aims to analyze a state-of-the-art text and image embedding neural network and guides through the application in the automotive domain. This approach enables the search for similar images and the search based on a human understandable text-based description. Our experiments show the automatability and generalizability of our proposed method for handling large data sets in the automotive domain.
Many analysis and prediction tasks require the extraction of structured data from unstructured texts. To solve it, this paper presents an end-to-end machine learning pipeline, Text2Struct, including a text annotation scheme, training data processing, and machine learning implementation. We formulated the mining problem as the extraction of metrics and units associated with numerals in the text. The Text2Struct was evaluated on an annotated text dataset collected from abstracts of medical publications regarding thrombectomy. In terms of prediction performance, a dice coefficient of 0.82 was achieved on the test dataset. By random sampling, most predicted relations between numerals and entities were well matched to the ground-truth annotations. These results show that the Text2Struct is viable for the mining of structured data from text without special templates or patterns. It is anticipated to further improve the pipeline by expanding the dataset and investigating other machine learning models. A code demonstration can be found at: https://github.com/zcc861007/CourseProject
Representing text into a multidimensional space can be done with sentence embedding models such as Sentence-BERT (SBERT). However, training these models when the data has a complex multilevel structure requires individually trained class-specific models, which increases time and computing costs. We propose a two step approach which enables us to map sentences according to their hierarchical memberships and polarity. At first we teach the upper level sentence space through an AdaCos loss function and then finetune with a novel loss function mainly based on the cosine similarity of intra-level pairs. We apply this method to three different datasets: two weakly supervised Big Five personality dataset obtained from English and Japanese Twitter data and the benchmark MNLI dataset. We show that our single model approach performs better than multiple class-specific classification models.
Accurate, real-time measurements of price index changes using electronic records are essential for tracking inflation and productivity in today's economic environment. We develop empirical hedonic models that can process large amounts of unstructured product data (text, images, prices, quantities) and output accurate hedonic price estimates and derived indices. To accomplish this, we generate abstract product attributes, or ``features,'' from text descriptions and images using deep neural networks, and then use these attributes to estimate the hedonic price function. Specifically, we convert textual information about the product to numeric features using large language models based on transformers, trained or fine-tuned using product descriptions, and convert the product image to numeric features using a residual network model. To produce the estimated hedonic price function, we again use a multi-task neural network trained to predict a product's price in all time periods simultaneously. To demonstrate the performance of this approach, we apply the models to Amazon's data for first-party apparel sales and estimate hedonic prices. The resulting models have high predictive accuracy, with $R^2$ ranging from $80\%$ to $90\%$. Finally, we construct the AI-based hedonic Fisher price index, chained at the year-over-year frequency. We contrast the index with the CPI and other electronic indices.
Despite the record-breaking performance in Text-to-Image (T2I) generation by Stable Diffusion, less research attention is paid to its adversarial robustness. In this work, we study the problem of adversarial attack generation for Stable Diffusion and ask if an adversarial text prompt can be obtained even in the absence of end-to-end model queries. We call the resulting problem 'query-free attack generation'. To resolve this problem, we show that the vulnerability of T2I models is rooted in the lack of robustness of text encoders, e.g., the CLIP text encoder used for attacking Stable Diffusion. Based on such insight, we propose both untargeted and targeted query-free attacks, where the former is built on the most influential dimensions in the text embedding space, which we call steerable key dimensions. By leveraging the proposed attacks, we empirically show that only a five-character perturbation to the text prompt is able to cause the significant content shift of synthesized images using Stable Diffusion. Moreover, we show that the proposed target attack can precisely steer the diffusion model to scrub the targeted image content without causing much change in untargeted image content.