How can we effectively generate missing data transformations among tables in a data repository? Multiple versions of the same tables are generated from the iterative process when data scientists and machine learning engineers fine-tune their ML pipelines, making incremental improvements. This process often involves data transformation and augmentation that produces an augmented table based on its base version and related tables. However, data transformations are often not well-documented or completely missing, resulting in poor traceability, reproducibility and explainability of ML pipelines. In this paper, we propose DATALoRE, a framework that explains data changes between an initial dataset and its augmented version to improves traceability. Given a base table, DATALoRE first discovers its potentially related tables from the data repository using a variety of data discovery techniques. DATALoRE then effectively leverages a large language model (LLM) to generate a variety of data transformations that lead to the augmented table. DATALoRE validates these transformations and selects the minimum number of related tables to ensure traceability and reproducibility of the ML pipelines. A preliminary experiment shows that DATALoRE is able to effectively recovery data transformations on two benchmark datasets.

Video is becoming a major part of contemporary data collection. It is increasingly important to process video selection queries --- selecting videos that contain target objects. Advances in neural networks allow us to detect the objects in an image, and thereby offer query systems to examine the content of the video. Unfortunately, neural network-based approaches have long inference times. Processing this type of query through a standard scan would be time-consuming and would involve applying complex detectors to numerous irrelevant videos. It is tempting to try to improve query times by computing an index in advance. But unfortunately, many frames will never be beneficial for any query. Time spent processing them, whether at index time or at query time, is simply wasted computation. We propose a novel index mechanism to optimize video selection queries with commonsense knowledge. Commonsense knowledge consists of fundamental information about the world, such as the fact that a tennis racket is a tool designed for hitting a tennis ball. To save computation, an inexpensive but lossy index can be intentionally created, but this may result in missed target objects and suboptimal query time performance. Our mechanism addresses this issue by constructing probabilistic models from commonsense knowledge to patch the lossy index and then prioritizing predicate-related videos at query time. This method can achieve significant performance improvements comparable to those of a full index while keeping the construction costs of a lossy index. We describe our prototype system, Paine, plus experiments on two video corpora. We show our best optimization method can process up to 97.79% fewer videos compared to baselines. Even the model constructed without any video content can yield a 75.39% improvement over baselines.

Data preparation is an essential step in every data-related effort, from scientific projects in academia to data-driven decision-making in industry. Typically, data preparation is not an interesting piece of a project --- it transforms raw data into a format that enables further innovative work. Because such scripts are never intended to be interesting, are project-specific, and are written in general-purpose languages, they can be tedious to understand and difficult to verify. As a result, data preparation scripts can easily become a breeding ground for poor engineering and statistical practices. Ideally, data preparation scripts are "admirably boring" --- they should serve the project, but otherwise be as simple and as standard as possible. We propose a bottom-up script standardization framework that takes a user's data preparation script and transforms it into a simpler, more standardized version of itself. Our framework takes the user's script not as an unchangeable definition of correctness, but as a sketch of the user's intent. We embedded this framework in a system called LucidScript.

Because video is becoming more popular and constitutes a major part of data collection, we have the need to process video selection queries --- selecting videos that contain target objects. However, a naïve scan of a video corpus without optimization would be extremely inefficient due to applying complex detectors to irrelevant videos. This demo presents Paine; a video query system that employs a novel index mechanism to optimize video selection queries via commonsense knowledge. Paine samples video frames to build an inexpensive lossy index, then leverages probabilistic models based on existing commonsense knowledge sources to capture the semantic-level correlation among video frames, thereby allowing Paine to predict the content of unindexed video. These models can predict which videos are likely to satisfy selection predicates so as to avoid Paine from processing irrelevant videos. We will demonstrate a system prototype of Paine for accelerating the processing of video selection queries, allowing VLDB'23 participants to use the Paine interface to run queries. Users can compare Paine with the baseline, the SCAN method.