Skip to yearly menu bar Skip to main content

Oral Session

Oral 4B Datasets & Benchmarks

La Nouvelle Orleans Ballroom A-C (level 2)
Wed 13 Dec 1:30 p.m. PST — 2:30 p.m. PST


Wed 13 Dec. 13:30 - 13:45 PST

Mesogeos: A multi-purpose dataset for data-driven wildfire modeling in the Mediterranean

Spyridon Kondylatos · Ioannis Prapas · Gustau Camps-Valls · Ioannis Papoutsis

We introduce Mesogeos, a large-scale multi-purpose dataset for wildfire modeling in the Mediterranean. Mesogeos integrates variables representing wildfire drivers (meteorology, vegetation, human activity) and historical records of wildfire ignitions and burned areas for 17 years (2006-2022). It is designed as a cloud-friendly spatio-temporal dataset, namely a datacube, harmonizing all variables in a grid of 1km x 1km x 1-day resolution. The datacube structure offers opportunities to assess machine learning (ML) usage in various wildfire modeling tasks. We extract two ML-ready datasets that establish distinct tracks to demonstrate this potential: (1) short-term wildfire danger forecasting and (2) final burned area estimation given the point of ignition. We define appropriate metrics and baselines to evaluate the performance of models in each track. By publishing the datacube, along with the code to create the ML datasets and models, we encourage the community to foster the implementation of additional tracks for mitigating the increasing threat of wildfires in the Mediterranean.

Wed 13 Dec. 13:45 - 14:00 PST

ClimSim: A large multi-scale dataset for hybrid physics-ML climate emulation

Sungduk Yu · Walter Hannah · Liran Peng · Jerry Lin · Mohamed Aziz Bhouri · Ritwik Gupta · Björn Lütjens · Justus C. Will · Gunnar Behrens · Julius Busecke · Nora Loose · Charles Stern · Tom Beucler · Bryce Harrop · Benjamin Hillman · Andrea Jenney · Savannah L. Ferretti · Nana Liu · Animashree Anandkumar · Noah Brenowitz · Veronika Eyring · Nicholas Geneva · Pierre Gentine · Stephan Mandt · Jaideep Pathak · Akshay Subramaniam · Carl Vondrick · Rose Yu · Laure Zanna · Tian Zheng · Ryan Abernathey · Fiaz Ahmed · David Bader · Pierre Baldi · Elizabeth Barnes · Christopher Bretherton · Peter Caldwell · Wayne Chuang · Yilun Han · YU HUANG · Fernando Iglesias-Suarez · Sanket Jantre · Karthik Kashinath · Marat Khairoutdinov · Thorsten Kurth · Nicholas Lutsko · Po-Lun Ma · Griffin Mooers · J. David Neelin · David Randall · Sara Shamekh · Mark Taylor · Nathan Urban · Janni Yuval · Guang Zhang · Mike Pritchard

Modern climate projections lack adequate spatial and temporal resolution due to computational constraints. A consequence is inaccurate and imprecise predictions of critical processes such as storms. Hybrid methods that combine physics with machine learning (ML) have introduced a new generation of higher fidelity climate simulators that can sidestep Moore's Law by outsourcing compute-hungry, short, high-resolution simulations to ML emulators. However, this hybrid ML-physics simulation approach requires domain-specific treatment and has been inaccessible to ML experts because of lack of training data and relevant, easy-to-use workflows. We present ClimSim, the largest-ever dataset designed for hybrid ML-physics research. It comprises multi-scale climate simulations, developed by a consortium of climate scientists and ML researchers. It consists of 5.7 billion pairs of multivariate input and output vectors that isolate the influence of locally-nested, high-resolution, high-fidelity physics on a host climate simulator's macro-scale physical state.The dataset is global in coverage, spans multiple years at high sampling frequency, and is designed such that resulting emulators are compatible with downstream coupling into operational climate simulators. We implement a range of deterministic and stochastic regression baselines to highlight the ML challenges and their scoring. The data ( and code ( are released openly to support the development of hybrid ML-physics and high-fidelity climate simulations for the benefit of science and society.

Wed 13 Dec. 14:00 - 14:15 PST

Quilt-1M: One Million Image-Text Pairs for Histopathology

Wisdom Ikezogwo · Saygin Seyfioglu · Fatemeh Ghezloo · Dylan Geva · Fatwir Sheikh Mohammed · Pavan Kumar Anand · Ranjay Krishna · Linda Shapiro

Recent accelerations in multi-modal applications have been made possible with the plethora of image and text data available online. However, the scarcity of analogous data in the medical field, specifically in histopathology, has slowed comparable progress. To enable similar representation learning for histopathology, we turn to YouTube, an untapped resource of videos, offering $1,087$ hours of valuable educational histopathology videos from expert clinicians.From YouTube, we curate QUILT: a large-scale vision-language dataset consisting of $802, 144$ image and text pairs.QUILT was automatically curated using a mixture of models, including large language models, handcrafted algorithms, human knowledge databases, and automatic speech recognition.In comparison, the most comprehensive datasets curated for histopathology amass only around $200$K samples.We combine QUILT with datasets from other sources, including Twitter, research papers, and the internet in general, to create an even larger dataset: QUILT-1M, with $1$M paired image-text samples, marking it as the largest vision-language histopathology dataset to date. We demonstrate the value of QUILT-1M by fine-tuning a pre-trained CLIP model. Our model outperforms state-of-the-art models on both zero-shot and linear probing tasks for classifying new histopathology images across $13$ diverse patch-level datasets of $8$ different sub-pathologies and cross-modal retrieval tasks.

Wed 13 Dec. 14:15 - 14:30 PST

BEDD: The MineRL BASALT Evaluation and Demonstrations Dataset for Training and Benchmarking Agents that Solve Fuzzy Tasks

Stephanie Milani · Anssi Kanervisto · Karolis Ramanauskas · Sander Schulhoff · Brandon Houghton · Rohin Shah

The MineRL BASALT competition has served to catalyze advances in learning from human feedback through four hard-to-specify tasks in Minecraft, such as create and photograph a waterfall. Given the completion of two years of BASALT competitions, we offer to the community a formalized benchmark through the BASALT Evaluation and Demonstrations Dataset (BEDD), which serves as a resource for algorithm development and performance assessment. BEDD consists of a collection of 26 million image-action pairs from nearly 14,000 videos of human players completing the BASALT tasks in Minecraft. It also includes over 3,000 dense pairwise human evaluations of human and algorithmic agents. These comparisons serve as a fixed, preliminary leaderboard for evaluating newly-developed algorithms. To enable this comparison, we present a streamlined codebase for benchmarking new algorithms against the leaderboard. In addition to presenting these datasets, we conduct a detailed analysis of the data from both datasets to guide algorithm development and evaluation. The released code and data are available at