Discovering tables from poorly maintained data lakes is a significant challenge in data management. Two key tasks are identifying joinable and unionable tables, crucial for data integration, analysis, and machine learning. However, there's a lack of a comprehensive benchmark for evaluating existing methods. To address this, we introduce LakeBench, a large-scale table discovery benchmark. It evaluates effectiveness, efficiency, and scalability of table join & union search methods. With over 16 million real tables, LakeBench is 1,600X larger than existing datasets and 100X larger in storage size. It includes synthesized and real queries with ground truth, totaling more than 10 thousand queries - 10X more than used in any existing evaluation. We spent over 7,500 human hours labeling these queries and constructing diverse query categories for thorough evaluation. Our benchmark thoroughly evaluates state-of-the-art table discovery methods, providing insights into their performance and highlighting research opportunities.

Coreset selection is a technique for efficient machine learning, which selects a subset of the training data to achieve similar model performance as using the full dataset. It can be performed with or without training machine learning models. Coreset selection with training, which iteratively trains the machine model and updates data items in the coreset, is time consuming. Coreset selection without training can select the coreset before training. Gradient approximation is the typical method, but it can also be slow when dealing with large training datasets as it requires multiple iterations and pairwise distance computations for each iteration. The state-of-the-art (SOTA) results w.r.t. effectiveness are achieved by the latter approach, i.e. gradient approximation. In this paper, we aim to significantly improve the efficiency of coreset selection while ensuring good effectiveness, by improving the SOTA approaches of using gradient descent without training machine learning models. Specifically, we present a highly efficient coreset selection framework that utilizes an approximation of the gradient. This is achieved by dividing the entire training set into multiple clusters, each of which contains items with similar feature distances (calculated using the Euclidean distance). Our framework further demonstrates that the full gradient can be bounded based on the maximum feature distance between each item and each cluster, allowing for more efficient coreset selection by iterating through these clusters. Additionally, we propose an efficient method for estimating the maximum feature distance using the product quantization technique. Our experiments on multiple real-world datasets demonstrate that we can improve the efficiency 3-10 times comparing with SOTA almost without sacrificing the accuracy.

Given a dataset with incomplete data (e.g., missing values), training a machine learning model over the incomplete data requires two steps. First, it requires a data-effective step that cleans the data in order to improve the data quality (and the model quality on the cleaned data). Second, it requires a data-efficient step that selects a core subset of the data (called coreset) such that the trained models on the entire data and the coreset have similar model quality, in order to save the computational cost of training. The first-data-effective-then-data-efficient methods are too costly, because they are expensive to clean the whole data; while the first-data-efficient-then-data-effective methods have low model quality, because they cannot select high-quality coreset for incomplete data. In this article, we investigate the problem of coreset selection over incomplete data for data-effective and data-efficient machine learning. The essential challenge is how to model the incomplete data for selecting high-quality coreset. To this end, we propose the GoodCore framework towards selecting a good coreset over incomplete data with low cost. To model the unknown complete data, we utilize the combinations of possible repairs as possible worlds of the incomplete data. Based on possible worlds, GoodCore selects an expected optimal coreset through gradient approximation without training ML models. We formally define the expected optimal coreset selection problem, prove its NP-hardness, and propose a greedy algorithm with an approximation ratio. To make GoodCore more efficient, we propose optimization methods that incorporate human-in-the-loop imputation or automatic imputation method into our framework. Moreover, a group-based strategy is utilized to further accelerate the coreset selection with incomplete data given large datasets. Experimental results show the effectiveness and efficiency of our framework with low cost.

While machine learning techniques, especially deep neural networks, have shown remarkable success in various applications, their performance is adversely affected by label errors in training data. Acquiring high-quality annotated data is both costly and time-consuming in real-world scenarios, requiring extensive human annotation and verification. Consequently, many industry-applied models are trained over data containing substantial noise, significantly degrading the performance of these models. To address this critical issue, we demonstrate IDE, a novel system that iteratively detects mislabeled instances and repairs the wrong labels. Specifically, IDE leverages the early loss observation and influence-based verification to iteratively identify mislabeled instances. When the mislabeled instances are obtained in each iteration, IDE will repair their labels to enhance detection accuracy for subsequent iterations. The framework automatically determines the termination point when the early loss is no longer effective. For uncertain instances, it generates pseudo labels to train a binary classification model, leveraging the model's generalization ability to make the final decision. With a real-life scenario, we demonstrate that IDE produces high-quality training data by effective mislabel detection and repair.

Using a large training dataset to train a big and powerful model -- a typical practice in modern deep learning, often suffers from two major problems: the expensive and slow training process and the error-prone labels. The existing approaches, targeting either speeding up the training by selecting a subset of representative training instances (subset selection) or eliminating the negative effect of mislabels during training (mislabel detection), do not perform well in this scenario due to overlooking one of these two problems. To fill this gap, we propose Deem, a novel data-efficient framework that selects a subset of representative training instances under label uncertainty. The key idea is to leverage the metadata produced during deep learning training, e.g., training losses and gradients, to estimate the label uncertainty and select the representative instances. In particular, we model the problem of subset selection under uncertainty as a problem of finding a subset that closely approximates the gradient of the whole training data set derived on soft labels. We show that it is an NP-hard problem with submodular property and propose a low complexity algorithm to solve this problem with an approximate ratio. Training on this small subset thus improves the training efficiency while guaranteeing the model's accuracy. Moreover, we propose an efficient strategy to dynamically refine this subset during the iterative training process. Extensive experiments on 6 datasets and 10 baselines demonstrate that Deem accelerates the training process up to 10X without sacrificing the model accuracy.