DEEM brings together researchers and practitioners at the intersection of applied machine learning, data management and systems research, with the goal to discuss the arising data management issues in ML application scenarios. The DEEM workshop will be held on Sunday, June 9th, in conjunction with SIGMOD/PODS 2024. The workshop will be held in-person.
The workshop solicits regular research papers (8 pages plus unlimited references) describing preliminary or completed research results, as well as short papers (up to 4 pages) such as reports on applications and tools or preliminary results, interesting use cases, problems, datasets, benchmarks, visionary ideas, and descriptions of system components and tools related to end-to-end ML pipelines. Submissions should follow the guidelines as for SIGMOD, i.e. use the sigconf template for the ACM proceedings format.
Follow us on twitter @deem_workshop or contact us via email at madelon@berkeley.edu. We also provide archived websites of previous versions of the workshop: DEEM 2017, DEEM 2018, DEEM 2019, DEEM 2020, DEEM 2021, DEEM 2022, and DEEM 2023.
DEEM 2024 Proceedings: https://dl.acm.org/doi/proceedings/10.1145/3650203
Artificial Intelligence (AI) is reshaping data-driven exploration. In this talk, we will explore how AutoML and data discovery enhance human capabilities. We present AlphaAutoML, an open-source Python library designed to support a wide range of machine learning tasks across various data types. AlphaAutoML combines deep reinforcement learning and meta-learning to effectively construct pipelines over a large collection of primitives. It seamlessly integrates AutoML within the data science lifecycle through an ecosystem of tools that facilitate user-in-the-loop tasks, such as selecting suitable pipelines and customizing these pipelines for complex problems. Additionally, we will discuss the emerging field of dataset search, a critical component of data-centric AI. We will review the opportunities it creates to enrich analytics and improve machine learning models, and present methods that support discovery in large dataset collections.
Applying Machine Learning (ML) in real-world scenarios is a challenging task. In recent years, the main focus of the data management community has been on creating systems and abstractions for the efficient training of ML models on large datasets. However, model training is only one of many steps in an end-to-end ML application, and a number of orthogonal data management problems arise from the large-scale use of ML and increased adoption large language models (LLMs).
For example, data preprocessing and feature extraction workloads may be complicated and require simultaneous execution of relational and linear algebraic operations. Next, model selection may involve searching many combinations of model architectures, features, and hyper-parameters to find the best-performing model. After model training, the resulting model may have to be deployed and integrated into business workflows and require lifecycle management using metadata and lineage. As a further complication, the resulting system may have to take into account a heterogeneous audience, ranging from domain experts without programming skills to data engineers and statisticians who develop custom algorithms. Many such challenges are human or engineer-centered (e.g., monitoring ML pipelines, leveraging LLMs for domain-specific tasks at scale), and DEEM uniquely encourages submissions in such topics.
Additionally, the importance of incorporating ethics and legal compliance into machine-assisted decision-making is being broadly recognized. Critical opportunities for improving data quality and representativeness, controlling for bias, and allowing humans to oversee and impact computational processes are missed if we do not consider the lifecycle stages upstream from model training and deployment. DEEM welcomes research on providing system-level support to data scientists who wish to develop and deploy responsible machine learning methods.
DEEM aims to bring together researchers and practitioners at the intersection of applied machine learning, data management and systems research, with the goal to discuss the arising data management issues in ML application scenarios.
We invite submissions in the following two tracks:
Madelon Hulsebos
Matteo Interlandi
Shreya Shankar