Collecting human judgements is currently the most reliable evaluation method for natural language generation systems. Automatic metrics have reported flaws when applied to measure quality aspects of generated text and have been shown to correlate poorly with human judgements. However, human evaluation is time and cost-intensive, and we lack consensus on designing and conducting human evaluation experiments. Thus there is a need for streamlined approaches for efficient collection of human judgements when evaluating natural language generation systems. Therefore, we present a dynamic approach to measure the required number of human annotations when evaluating generated outputs in relative comparison settings. We propose an agent-based framework of human evaluation to assess multiple labelling strategies and methods to decide the better model in a simulation and a crowdsourcing case study. The main results indicate that a decision about the superior model can be made with high probability across different labelling strategies, where assigning a single random worker per task requires the least overall labelling effort and thus the least cost.
Deep learning based models have dominated the current landscape of production recommender systems. Furthermore, recent years have witnessed an exponential growth of the model scale--from Google's 2016 model with 1 billion parameters to the latest Facebook's model with 12 trillion parameters. Significant quality boost has come with each jump of the model capacity, which makes us believe the era of 100 trillion parameters is around the corner. However, the training of such models is challenging even within industrial scale data centers. This difficulty is inherited from the staggering heterogeneity of the training computation--the model's embedding layer could include more than 99.99% of the total model size, which is extremely memory-intensive; while the rest neural network is increasingly computation-intensive. To support the training of such huge models, an efficient distributed training system is in urgent need. In this paper, we resolve this challenge by careful co-design of both the optimization algorithm and the distributed system architecture. Specifically, in order to ensure both the training efficiency and the training accuracy, we design a novel hybrid training algorithm, where the embedding layer and the dense neural network are handled by different synchronization mechanisms; then we build a system called Persia (short for parallel recommendation training system with hybrid acceleration) to support this hybrid training algorithm. Both theoretical demonstration and empirical study up to 100 trillion parameters have conducted to justified the system design and implementation of Persia. We make Persia publicly available (at https://github.com/PersiaML/Persia) so that anyone would be able to easily train a recommender model at the scale of 100 trillion parameters.
Deep learning based models have dominated the current landscape of production recommender systems. Furthermore, recent years have witnessed an exponential growth of the model scale--from Google's 2016 model with 1 billion parameters to the latest Facebook's model with 12 trillion parameters. Significant quality boost has come with each jump of the model capacity, which makes us believe the era of 100 trillion parameters is around the corner. However, the training of such models is challenging even within industrial scale data centers. This difficulty is inherited from the staggering heterogeneity of the training computation--the model's embedding layer could include more than 99.99% of the total model size, which is extremely memory-intensive; while the rest neural network is increasingly computation-intensive. To support the training of such huge models, an efficient distributed training system is in urgent need. In this paper, we resolve this challenge by careful co-design of both the optimization algorithm and the distributed system architecture. Specifically, in order to ensure both the training efficiency and the training accuracy, we design a novel hybrid training algorithm, where the embedding layer and the dense neural network are handled by different synchronization mechanisms; then we build a system called Persia (short for parallel recommendation training system with hybrid acceleration) to support this hybrid training algorithm. Both theoretical demonstration and empirical study up to 100 trillion parameters have conducted to justified the system design and implementation of Persia. We make Persia publicly available (at https://github.com/PersiaML/Persia) so that anyone would be able to easily train a recommender model at the scale of 100 trillion parameters.
Semantic segmentation of fine-resolution urban scene images plays a vital role in extensive practical applications, such as land cover mapping, urban change detection, environmental protection and economic assessment. Driven by rapid developments in deep learning technologies, the convolutional neural network (CNN) has dominated the semantic segmentation task for many years. Convolutional neural networks adopt hierarchical feature representation, demonstrating strong local information extraction. However, the local property of the convolution layer limits the network from capturing global context that is crucial for precise segmentation. Recently, Transformer comprise a hot topic in the computer vision domain. Transformer demonstrates the great capability of global information modelling, boosting many vision tasks, such as image classification, object detection and especially semantic segmentation. In this paper, we propose an efficient hybrid Transformer (EHT) for real-time urban scene segmentation. The EHT adopts a hybrid structure with and CNN-based encoder and a transformer-based decoder, learning global-local context with lower computation. Extensive experiments demonstrate that our EHT has faster inference speed with competitive accuracy compared with state-of-the-art lightweight models. Specifically, the proposed EHT achieves a 66.9% mIoU on the UAVid test set and outperforms other benchmark networks significantly. The code will be available soon.
With the recent surge in social applications relying on knowledge graphs, the need for techniques to ensure fairness in KG based methods is becoming increasingly evident. Previous works have demonstrated that KGs are prone to various social biases, and have proposed multiple methods for debiasing them. However, in such studies, the focus has been on debiasing techniques, while the relations to be debiased are specified manually by the user. As manual specification is itself susceptible to human cognitive bias, there is a need for a system capable of quantifying and exposing biases, that can support more informed decisions on what to debias. To address this gap in the literature, we describe a framework for identifying biases present in knowledge graph embeddings, based on numerical bias metrics. We illustrate the framework with three different bias measures on the task of profession prediction, and it can be flexibly extended to further bias definitions and applications. The relations flagged as biased can then be handed to decision makers for judgement upon subsequent debiasing.
Semantic segmentation of fine-resolution urban scene images plays a vital role in extensive practical applications, such as land cover mapping, urban change detection, environmental protection and economic assessment. Driven by rapid developments in deep learning technologies, convolutional neural networks (CNNs) have dominated the semantic segmentation task for many years. Convolutional neural networks adopt hierarchical feature representation and have strong local context extraction. However, the local property of the convolution layer limits the network from capturing global information that is crucial for improving fine-resolution image segmentation. Recently, Transformer comprise a hot topic in the computer vision domain. Vision Transformer demonstrates the great capability of global information modelling, boosting many vision tasks, such as image classification, object detection and especially semantic segmentation. In this paper, we propose an efficient hybrid Transformer (EHT) for semantic segmentation of urban scene images. EHT takes advantage of CNNs and Transformer, learning global-local context to strengthen the feature representation. Extensive experiments demonstrate that EHT has higher efficiency with competitive accuracy compared with state-of-the-art benchmark methods. Specifically, the proposed EHT achieves a 67.0% mIoU on the UAVid test set and outperforms other lightweight models significantly. The code will be available soon.
The Bayes error rate (BER) is a fundamental concept in machine learning that quantifies the best possible accuracy any classifier can achieve on a fixed probability distribution. Despite years of research on building estimators of lower and upper bounds for the BER, these were usually compared only on synthetic datasets with known probability distributions, leaving two key questions unanswered: (1) How well do they perform on real-world datasets?, and (2) How practical are they? Answering these is not trivial. Apart from the obvious challenge of an unknown BER for real-world datasets, there are two main aspects any BER estimator needs to overcome in order to be applicable in real-world settings: (1) the computational and sample complexity, and (2) the sensitivity and selection of hyper-parameters. In this work, we propose FeeBee, the first principled framework for analyzing and comparing BER estimators on any modern real-world dataset with unknown probability distribution. We achieve this by injecting a controlled amount of label noise and performing multiple evaluations on a series of different noise levels, supported by a theoretical result which allows drawing conclusions about the evolution of the BER. By implementing and analyzing 7 multi-class BER estimators on 6 commonly used datasets of the computer vision and NLP domains, FeeBee allows a thorough study of these estimators, clearly identifying strengths and weaknesses of each, whilst being easily deployable on any future BER estimator.
Graph structured data have enabled several successful applications such as recommendation systems and traffic prediction, given the rich node features and edges information. However, these high-dimensional features and high-order adjacency information are usually heterogeneous and held by different data holders in practice. Given such vertical data partition (e.g., one data holder will only own either the node features or edge information), different data holders have to develop efficient joint training protocols rather than directly transfer data to each other due to privacy concerns. In this paper, we focus on the edge privacy, and consider a training scenario where Bob with node features will first send training node features to Alice who owns the adjacency information. Alice will then train a graph neural network (GNN) with the joint information and release an inference API. During inference, Bob is able to provide test node features and query the API to obtain the predictions for test nodes. Under this setting, we first propose a privacy attack LinkTeller via influence analysis to infer the private edge information held by Alice via designing adversarial queries for Bob. We then empirically show that LinkTeller is able to recover a significant amount of private edges, outperforming existing baselines. To further evaluate the privacy leakage, we adapt an existing algorithm for differentially private graph convolutional network (DP GCN) training and propose a new DP GCN mechanism LapGraph. We show that these DP GCN mechanisms are not always resilient against LinkTeller empirically under mild privacy guarantees ($\varepsilon>5$). Our studies will shed light on future research towards designing more resilient privacy-preserving GCN models; in the meantime, provide an in-depth understanding of the tradeoff between GCN model utility and robustness against potential privacy attacks.
Forest carbon offsets are increasingly popular and can play a significant role in financing climate mitigation, forest conservation, and reforestation. Measuring how much carbon is stored in forests is, however, still largely done via expensive, time-consuming, and sometimes unaccountable field measurements. To overcome these limitations, many verification bodies are leveraging machine learning (ML) algorithms to estimate forest carbon from satellite or aerial imagery. Aerial imagery allows for tree species or family classification, which improves the satellite imagery-based forest type classification. However, aerial imagery is significantly more expensive to collect and it is unclear by how much the higher resolution improves the forest carbon estimation. This proposal paper describes the first systematic comparison of forest carbon estimation from aerial imagery, satellite imagery, and ground-truth field measurements via deep learning-based algorithms for a tropical reforestation project. Our initial results show that forest carbon estimates from satellite imagery can overestimate above-ground biomass by up to 10-times for tropical reforestation projects. The significant difference between aerial and satellite-derived forest carbon measurements shows the potential for aerial imagery-based ML algorithms and raises the importance to extend this study to a global benchmark between options for carbon measurements.