Model hubs with many pre-trained models (PTMs) have become a cornerstone of deep learning. Although built at a high cost, they remain \emph{under-exploited} -- practitioners usually pick one PTM from the provided model hub by popularity and then fine-tune the PTM to solve the target task. This na\"ive but common practice poses two obstacles to full exploitation of pre-trained model hubs: first, the PTM selection by popularity has no optimality guarantee, and second, only one PTM is used while the remaining PTMs are ignored. An alternative might be to consider all possible combinations of PTMs and extensively fine-tune each combination, but this would not only be prohibitive computationally but may also lead to statistical over-fitting. In this paper, we propose a new paradigm for exploiting model hubs that is intermediate between these extremes. The paradigm is characterized by two aspects: (1) We use an evidence maximization procedure to estimate the maximum value of label evidence given features extracted by pre-trained models. This procedure can rank all the PTMs in a model hub for various types of PTMs and tasks \emph{before fine-tuning}. (2) The best ranked PTM can either be fine-tuned and deployed if we have no preference for the model's architecture or the target PTM can be tuned by the top $K$ ranked PTMs via a Bayesian procedure that we propose. This procedure, which we refer to as \emph{B-Tuning}, not only improves upon specialized methods designed for tuning homogeneous PTMs, but also applies to the challenging problem of tuning heterogeneous PTMs where it yields a new level of benchmark performance.

Parkinson's disease (PD) is a progressive neurodegenerative disorder that results in a variety of motor dysfunction symptoms, including tremors, bradykinesia, rigidity and postural instability. The diagnosis of PD mainly relies on clinical experience rather than a definite medical test, and the diagnostic accuracy is only about 73-84% since it is challenged by the subjective opinions or experiences of different medical experts. Therefore, an efficient and interpretable automatic PD diagnosis system is valuable for supporting clinicians with more robust diagnostic decision-making. To this end, we propose to classify Parkinson's tremor since it is one of the most predominant symptoms of PD with strong generalizability. Different from other computer-aided time and resource-consuming Parkinson's Tremor (PT) classification systems that rely on wearable sensors, we propose SPAPNet, which only requires consumer-grade non-intrusive video recording of camera-facing human movements as input to provide undiagnosed patients with low-cost PT classification results as a PD warning sign. For the first time, we propose to use a novel attention module with a lightweight pyramidal channel-squeezing-fusion architecture to extract relevant PT information and filter the noise efficiently. This design aids in improving both classification performance and system interpretability. Experimental results show that our system outperforms state-of-the-arts by achieving a balanced accuracy of 90.9% and an F1-score of 90.6% in classifying PT with the non-PT class.

Building document-grounded dialogue systems have received growing interest as documents convey a wealth of human knowledge and commonly exist in enterprises. Wherein, how to comprehend and retrieve information from documents is a challenging research problem. Previous work ignores the visual property of documents and treats them as plain text, resulting in incomplete modality. In this paper, we propose a Layout-aware document-level Information Extraction dataset, LIE, to facilitate the study of extracting both structural and semantic knowledge from visually rich documents (VRDs), so as to generate accurate responses in dialogue systems. LIE contains 62k annotations of three extraction tasks from 4,061 pages in product and official documents, becoming the largest VRD-based information extraction dataset to the best of our knowledge. We also develop benchmark methods that extend the token-based language model to consider layout features like humans. Empirical results show that layout is critical for VRD-based extraction, and system demonstration also verifies that the extracted knowledge can help locate the answers that users care about.

Despite a growing recognition of the importance of insurgent group structure on conflict outcomes, there is very little empirical research thereon. Though this problem is rooted in the inaccessibility of data on militant group structure, insurgents frequently publish large volumes of image data on the internet. In this paper, I develop a new methodology that leverages this abundant but underutilized source of data by automating the creation of a social network graph based on co-appearance in photographs using deep learning. Using a trove of 19,115 obituary images published online by the PKK, a Kurdish militant group in Turkey, I demonstrate that an individual's centrality in the resulting co-appearance network is closely correlated with their rank in the insurgent group.

This study discusses the effect of semi-supervised learning in combination with pretrained language models for data-to-text generation. It is not known whether semi-supervised learning is still helpful when a large-scale language model is also supplemented. This study aims to answer this question by comparing a data-to-text system only supplemented with a language model, to two data-to-text systems that are additionally enriched by a data augmentation or a pseudo-labeling semi-supervised learning approach. Results show that semi-supervised learning results in higher scores on diversity metrics. In terms of output quality, extending the training set of a data-to-text system with a language model using the pseudo-labeling approach did increase text quality scores, but the data augmentation approach yielded similar scores to the system without training set extension. These results indicate that semi-supervised learning approaches can bolster output quality and diversity, even when a language model is also present.

As fishermen have noticed this behaviour, they have used both natural and man-made floating objects, or drifting Fish Aggregating Devices (dFADs), as a tool for finding and catching tropical tunas. The use of dFADs in tuna purse-seine fisheries has gradually increased since the 1980s to the present time, where vessels using dFADs now contribute to 36% of the world's total tropical tuna catch (Davies et al., 2014; Wain et al., 2021; ISSF, 2021). These widespread changes have highlighted the need to better understand the potential ecological effects of dFADs on tuna ecology and the marine environment, in order to ensure adequate management of fish stocks and dFAD usage. Indeed, both the dynamics of how and why tuna associate to dFADs are still poorly understood. Regarding the reasons behind tuna aggregation to dFADs, a number of hypotheses have been suggested (Fréon and Dagorn, 2000; Dempster and Taquet, 2004; Castro et al., 2002). Of these, two have gained traction: the "meeting-point" hypothesis, which considers that dFADs facilitate the encounter between individuals or schools, thus constituting larger schools that could benefit survival rates (Castro et al., 2002); and the "indicator-log" hypothesis, by which tunas may be safeguarding the survival of their eggs, larvae and juvenile stages by using drifting objects as indicators of areas where plankton and food is readily available (Hall et al., 1992). This scenario has led some authors to postulate that man-made dFADs could have detrimental effects on tuna populations by creating a so-called "ecological trap" which would lead tuna to remain associated to dFADs even as these drift into areas that could negatively affect the tuna's behaviour and biology (Marsac et al., 2000; Hallier and Gaertner, 2008). To the best of our knowledge, there is yet no sufficient evidence to either confirm or reject this hypothesis (see Dagorn et al. (2012) and references therein). Given the concerns around the widespread use of dFADs in tuna fisheries today, it is not surprising that a considerable amount of research has been devoted to characterizing the dynamics at play when tunas aggregate to dFADs.

Attention-based models have been widely used in many areas, such as computer vision and natural language processing. However, relevant applications in time series classification (TSC) have not been explored deeply yet, causing a significant number of TSC algorithms still suffer from general problems of attention mechanism, like quadratic complexity. In this paper, we promote the efficiency and performance of the attention mechanism by proposing our flexible multi-head linear attention (FMLA), which enhances locality awareness by layer-wise interactions with deformable convolutional blocks and online knowledge distillation. What's more, we propose a simple but effective mask mechanism that helps reduce the noise influence in time series and decrease the redundancy of the proposed FMLA by masking some positions of each given series proportionally. To stabilize this mechanism, samples are forwarded through the model with random mask layers several times and their outputs are aggregated to teach the same model with regular mask layers. We conduct extensive experiments on 85 UCR2018 datasets to compare our algorithm with 11 well-known ones and the results show that our algorithm has comparable performance in terms of top-1 accuracy. We also compare our model with three Transformer-based models with respect to the floating-point operations per second and number of parameters and find that our algorithm achieves significantly better efficiency with lower complexity.

In past years, the world has switched to many-core and multi-core shared memory architectures. As a result, there is a growing need to utilize these architectures by introducing shared memory parallelization schemes to software applications. OpenMP is the most comprehensive API that implements such schemes, characterized by a readable interface. Nevertheless, introducing OpenMP into code is challenging due to pervasive pitfalls in management of parallel shared memory. To facilitate the performance of this task, many source-to-source (S2S) compilers have been created over the years, tasked with inserting OpenMP directives into code automatically. In addition to having limited robustness to their input format, these compilers still do not achieve satisfactory coverage and precision in locating parallelizable code and generating appropriate directives. In this work, we propose leveraging recent advances in ML techniques, specifically in natural language processing (NLP), to replace S2S compilers altogether. We create a database (corpus), Open-OMP, specifically for this goal. Open-OMP contains over 28,000 code snippets, half of which contain OpenMP directives while the other half do not need parallelization at all with high probability. We use the corpus to train systems to automatically classify code segments in need of parallelization, as well as suggest individual OpenMP clauses. We train several transformer models, named PragFormer, for these tasks, and show that they outperform statistically-trained baselines and automatic S2S parallelization compilers in both classifying the overall need for an OpenMP directive and the introduction of private and reduction clauses. Our source code and database are available at: https://github.com/Scientific-Computing-Lab-NRCN/PragFormer.

Automatic Speech Recognition (ASR) is a key element in new services that helps users to interact with an automated system. Deep learning methods have made it possible to deploy systems with word error rates below 5% for ASR of English. However, the use of these methods is only available for languages with hundreds or thousands of hours of audio and their corresponding transcriptions. For the so-called low-resource languages to speed up the availability of resources that can improve the performance of their ASR systems, methods of creating new resources on the basis of existing ones are being investigated. In this paper we describe our data augmentation approach to improve the results of ASR models for low-resource and agglutinative languages. We carry out experiments developing an ASR for Quechua using the wav2letter++ model. We reduced WER by 8.73% through our approach to the base model. The resulting ASR model obtained 22.75% WER and was trained with 99 hours of original resources and 99 hours of synthetic data obtained with a combination of text augmentation and synthetic speech generati

The pre-training of large language models usually requires massive amounts of resources, both in terms of computation and data. Frequently used web sources such as Common Crawl might contain enough noise to make this pre-training sub-optimal. In this work, we experiment with different sampling methods from the Spanish version of mC4, and present a novel data-centric technique which we name $\textit{perplexity sampling}$ that enables the pre-training of language models in roughly half the amount of steps and using one fifth of the data. The resulting models are comparable to the current state-of-the-art, and even achieve better results for certain tasks. Our work is proof of the versatility of Transformers, and paves the way for small teams to train their models on a limited budget. Our models are available at this $\href{https://huggingface.co/bertin-project}{URL}$.