Software

When meta-research (research on research) makes an observation or points out a problem (such as a flaw in methodology), the project should be repeated later to determine whether the problem remains. For this we need meta-research that is reproducible and updatable, or living meta-research. In this talk, we introduce the concept of living meta-research, examine prequels to this idea, and point towards standards and technologies that could assist researchers in doing living meta-research. We introduce technologies like natural language processing, which can help with automation of meta-research, which in turn will make the research easier to reproduce/update. Further, we showcase our open-source litmining ecosystem, which includes pubget (for downloading full-text journal articles), labelbuddy (for manually extracting information), and pubextract (for automatically extracting information). With these tools, you can simplify the tedious data collection and information extraction steps in meta-research, and then focus on analyzing the text. We will then describe some living meta-research projects to illustrate the use of these tools. For example, we’ll show how we used GPT along with our tools to extract information about study participants. Essentially, this talk will introduce you to the concept of meta-research, some tools for doing meta-research, and some examples. Particularly, we want you to take away the fact that there are many interesting open questions in meta-research, and you can easily learn the tools to answer them. Check out our tools at https://litmining.github.io/

The abstract representation of space has been extensively studied in the hippocampus and entorhinal cortex. Yet while a growing variety of theoretical models have been proposed to capture the rich neural and behavioral phenomenology associated with these regions, it remains difficult to compare these theories systematically against each other using the full range of empirical data. To address this gap, we built an open-source standardised software framework, named NeuralPlayground, to facilitate comparisons of hippocampus and entorhinal cortex models. This Python software package offers a reproducible way to compare models consistently against a centralised library of published experimental results, including neural recordings and animal behavior. The framework currently contains implementations of three Agents, including a excitatory/inhibitory plasticity agent and the Tolman-Eichenbaum machine (TEM), three Experiments providing simple interfaces to publicly available neural and behavioral data; a customizable 2-dimensional Arena (continuous and discrete) able to produce common experimental environments such as T-maze, circular and dynamical arenas which the agents can move in and interact with, and a Comparison tool to facilitate systematic comparisons by enabling users to pick from any Agent, Arena, Experiments, and metrics to produce a comparison of the results between artificial agents and experimental measurements. As a result, we hope to generate new insights into the model theory and implementation, illuminating the strengths and weaknesses of each model. We hope our framework, available at github.com/anonymous, offers a foundation that the community will build upon, toward a shared computational understanding of the hippocampus and entorhinal cortex.

Nick Kuhn joins Daniel Whitenack and Chris Benson to discuss deploying AI agents alongside conventional enterprise applications. Topics include agent build packs, MCP gateways, shared memory, identity, sandboxing and lessons from platform engineering.

Support for focused projects that strengthen the R ecosystem’s technical or social infrastructure through open-source software and tools. Proposals should benefit the wider R community, identify clear deliverables and milestones, and show preliminary work that reduces delivery risk. Narrow-user projects, highly exploratory research, hardware/cloud purchases and conference/workshop sponsorship are generally outside this call. The second 2026 round is open from 1 September through 1 October 2026 at 11:59 p.m. US Eastern Time. Submit a two-to-five-page proposal using the required template through the linked Google-account form. Funding is paid 50% upon contract signature and 50% on completion; no fixed award cap is published.

Hamza Tahir discusses how experience with MLOps can inform more reliable agent systems. The conversation covers workflow execution, replayability and observability, including the design of agent harnesses and how ZenML's Kitaru approaches the practical work of running agents in production.

Chris Benson and Daniel Whitenack examine a reported security incident involving OpenAI agents and Hugging Face. They reconstruct the episode's account of what happened and discuss what it suggests about sandboxing, network access and the controls needed when AI systems can execute actions.