Textual knowledge bases such as Wikipedia require considerable effort to keep up to date and consistent. While automated writing assistants could potentially ease this burden, the problem of suggesting edits grounded in external knowledge has been under-explored. In this paper, we introduce the novel generation task of *faithfully reflecting updated information in text*(FRUIT) where the goal is to update an existing article given new evidence. We release the FRUIT-WIKI dataset, a collection of over 170K distantly supervised data produced from pairs of Wikipedia snapshots, along with our data generation pipeline and a gold evaluation set of 914 instances whose edits are guaranteed to be supported by the evidence. We provide benchmark results for popular generation systems as well as EDIT5 -- a T5-based approach tailored to editing we introduce that establishes the state of the art. Our analysis shows that developing models that can update articles faithfully requires new capabilities for neural generation models, and opens doors to many new applications.
Recent advances in contrastive representation learning over paired image-text data have led to models such as CLIP that achieve state-of-the-art performance for zero-shot classification and distributional robustness. Such models typically require joint reasoning in the image and text representation spaces for downstream inference tasks. Contrary to prior beliefs, we demonstrate that the image and text representations learned via a standard contrastive objective are not interchangeable and can lead to inconsistent downstream predictions. To mitigate this issue, we formalize consistency and propose CyCLIP, a framework for contrastive representation learning that explicitly optimizes for the learned representations to be geometrically consistent in the image and text space. In particular, we show that consistent representations can be learned by explicitly symmetrizing (a) the similarity between the two mismatched image-text pairs (cross-modal consistency); and (b) the similarity between the image-image pair and the text-text pair (in-modal consistency). Empirically, we show that the improved consistency in CyCLIP translates to significant gains over CLIP, with gains ranging from 10%-24% for zero-shot classification accuracy on standard benchmarks (CIFAR-10, CIFAR-100, ImageNet1K) and 10%-27% for robustness to various natural distribution shifts. The code is available at https://github.com/goel-shashank/CyCLIP.
The past few years has seen the application of machine learning utilised in the exploration of new materials. As in many fields of research - the vast majority of knowledge is published as text, which poses challenges in either a consolidated or statistical analysis across studies and reports. Such challenges include the inability to extract quantitative information, and in accessing the breadth of non-numerical information. To address this issue, the application of natural language processing (NLP) has been explored in several studies to date. In NLP, assignment of high-dimensional vectors, known as embeddings, to passages of text preserves the syntactic and semantic relationship between words. Embeddings rely on machine learning algorithms and in the present work, we have employed the Word2Vec model, previously explored by others, and the BERT model - applying them towards a unique challenge in materials engineering. That challenge is the search for chromate replacements in the field of corrosion protection. From a database of over 80 million records, a down-selection of 5990 papers focused on the topic of corrosion protection were examined using NLP. This study demonstrates it is possible to extract knowledge from the automated interpretation of the scientific literature and achieve expert human level insights.
Polarization and echo chambers are often studied in the context of explicitly political events such as elections, and little scholarship has examined the mixing of political groups in non-political contexts. A major obstacle to studying political polarization in non-political contexts is that political leaning (i.e., left vs right orientation) is often unknown. Nonetheless, political leaning is known to correlate (sometimes quite strongly) with many lifestyle choices leading to stereotypes such as the "latte-drinking liberal." We develop a machine learning classifier to infer political leaning from non-political text and, optionally, the accounts a user follows on social media. We use Voter Advice Application results shared on Twitter as our groundtruth and train and test our classifier on a Twitter dataset comprising the 3,200 most recent tweets of each user after removing any tweets with political text. We correctly classify the political leaning of most users (F1 scores range from 0.70 to 0.85 depending on coverage). We find no relationship between the level of political activity and our classification results. We apply our classifier to a case study of news sharing in the UK and discover that, in general, the sharing of political news exhibits a distinctive left-right divide while sports news does not.
This paper is on soft prompt learning for Vision \& Language (V&L) models. Similarly to their NLP counterparts, V\&L models can be adapted to a downstream task by learning soft continuous prompts using a few training examples. Current methods learn the soft prompts by minimizing a cross-entropy loss using as class weights the features obtained by passing the prompts plus the class names through the text encoder. Such methods, however, significantly overfit the training data suffering from large accuracy degradation when tested on unseen classes from the same domain. Our main contribution, in this paper, is a surprisingly simple approach to alleviate this problem: we use a second cross entropy loss to minimize the distance between the learned soft prompts and a set of hand-engineered manual prompts (obtained by prompt engineering). The proposed loss can be interpreted in multiple ways including as a regularizer, as a means for language-based augmentation, and as a way of learning more discriminative class centroids. Importantly, our formulation is inherently amenable to including, during training, virtual classes, i.e. class names for which no visual samples are available, further increasing the robustness of the learned prompts. Through extensive evaluations on 11 datasets, we show that our approach (a) significantly outperforms all prior works on soft prompting, and (b) matches and surpasses, for the first time, the accuracy on novel classes obtained by hand-crafted prompts and CLIP for the majority of the test datasets. Code will be made available.
Although current state-of-the-art language models have achieved impressive results in numerous natural language processing tasks, still they could not solve the problem of producing repetitive, dull and sometimes inconsistent text in open-ended text generation. Studies often attribute this problem to the maximum likelihood training objective, and propose alternative approaches by using stochastic decoding methods or altering the training objective. However, there is still a lack of consistent evaluation metrics to directly compare the efficacy of these solutions. In this work, we study different evaluation metrics that have been proposed to evaluate quality, diversity and consistency of machine-generated text. From there, we propose a practical pipeline to evaluate language models in open-ended generation task, and research on how to improve the model's performance in all dimensions by leveraging different auxiliary training objectives.
Recently, leveraging BERT pre-training to improve the phoneme encoder in text to speech (TTS) has drawn increasing attention. However, the works apply pre-training with character-based units to enhance the TTS phoneme encoder, which is inconsistent with the TTS fine-tuning that takes phonemes as input. Pre-training only with phonemes as input can alleviate the input mismatch but lack the ability to model rich representations and semantic information due to limited phoneme vocabulary. In this paper, we propose MixedPhoneme BERT, a novel variant of the BERT model that uses mixed phoneme and sup-phoneme representations to enhance the learning capability. Specifically, we merge the adjacent phonemes into sup-phonemes and combine the phoneme sequence and the merged sup-phoneme sequence as the model input, which can enhance the model capacity to learn rich contextual representations. Experiment results demonstrate that our proposed Mixed-Phoneme BERT significantly improves the TTS performance with 0.30 CMOS gain compared with the FastSpeech 2 baseline. The Mixed-Phoneme BERT achieves 3x inference speedup and similar voice quality to the previous TTS pre-trained model PnG BERT
In this work, we present a new dataset and a computational strategy for a digital coach that aims to guide users in practicing the protocols of self-attachment therapy. Our framework augments a rule-based conversational agent with a deep-learning classifier for identifying the underlying emotion in a user's text response, as well as a deep-learning assisted retrieval method for producing novel, fluent and empathetic utterances. We also craft a set of human-like personas that users can choose to interact with. Our goal is to achieve a high level of engagement during virtual therapy sessions. We evaluate the effectiveness of our framework in a non-clinical trial with N=16 participants, all of whom have had at least four interactions with the agent over the course of five days. We find that our platform is consistently rated higher for empathy, user engagement and usefulness than the simple rule-based framework. Finally, we provide guidelines to further improve the design and performance of the application, in accordance with the feedback received.
Word Error Rate (WER) is the primary metric used to assess automatic speech recognition (ASR) model quality. It has been shown that ASR models tend to have much higher WER on speakers with speech impairments than typical English speakers. It is hard to determine if models can be be useful at such high error rates. This study investigates the use of BERTScore, an evaluation metric for text generation, to provide a more informative measure of ASR model quality and usefulness. Both BERTScore and WER were compared to prediction errors manually annotated by Speech Language Pathologists for error type and assessment. BERTScore was found to be more correlated with human assessment of error type and assessment. BERTScore was specifically more robust to orthographic changes (contraction and normalization errors) where meaning was preserved. Furthermore, BERTScore was a better fit of error assessment than WER, as measured using an ordinal logistic regression and the Akaike's Information Criterion (AIC). Overall, our findings suggest that BERTScore can complement WER when assessing ASR model performance from a practical perspective, especially for accessibility applications where models are useful even at lower accuracy than for typical speech.
The Persian language is an inflectional SOV language. This fact makes Persian a more uncertain language. However, using techniques such as ZWNJ recognition, punctuation restoration, and Persian Ezafe construction will lead us to a more understandable and precise language. In most of the works in Persian, these techniques are addressed individually. Despite that, we believe that for text refinement in Persian, all of these tasks are necessary. In this work, we proposed a ViraPart framework that uses embedded ParsBERT in its core for text clarifications. First, used the BERT variant for Persian following by a classifier layer for classification procedures. Next, we combined models outputs to output cleartext. In the end, the proposed model for ZWNJ recognition, punctuation restoration, and Persian Ezafe construction performs the averaged F1 macro scores of 96.90%, 92.13%, and 98.50%, respectively. Experimental results show that our proposed approach is very effective in text refinement for the Persian language.