Name: The State of AI in Large Scale Automated Refactoring
Start: 2025-01-21T23:45:00.000Z
End: 2025-01-22T01:45:00.000Z
Location: Roam

**Abstract:**
Imagine your Java application starting up in just five milliseconds. Sounds incredible, right? Almost too good to be true? And no, this isn’t about native images. There's an innovative solution called Coordinated Restore at Checkpoint (CRaC).

This concept emerged initially from the Linux CRIU project: Checkpoint/Restore In Userspace. However, CRIU doesn't always deliver reliable results. That’s where CRaC comes in, offering a JVM-specific implementation.

In this talk, we’ll examine the complexities of CRIU and then turn to CRaC. We will discuss how CRaC operates within the Java ecosystem, its integration with Spring, and how you can use it in your applications for near-instant startup times.

From this talk you will learn what is possible to CRaC and what is not, what are the benefits and limitations, and what is the cost of CRaC-ing.

**ANNOUNCEMENT: DEVNEXUS (MARCH 4-6, 2025) TICKET SALE IS NOW OPEN**

Head over to [devnexus.com](https://devnexus.com/) and grab a two-day pass for only $545

Atlanta Java Users Group

Vincent Mayers

Glenn Renfro

சக்திவேல் Sakthivel

CRaCing Java Snapshots

Roam

It's that time of year again when we are incredibly fortunate to host the amazing Dr. Venkat Subramaniam at our monthly meeting. Venkat is a published author, educator, trainer, internationally renowned speaker, and Java Champion. But more importantly, you could not meet a more caring and wonderful person.

Join us as Venkat shares his unique and expert knowledge on all things software architecture and development. Don't miss this opportunity to meet and learn from a true software craftsman

Eric Cobb

An Audience with Dr. Venkat Subramaniam

LLMs are data-hungry, and when it comes to source code, the text simply isn't enough to make large-scale inferences about a codebase. As we know, code has a unique structure and strict grammar, as well as dependencies and type information that must be deterministically resolved by a compiler. This information could be incredibly useful for AI but is invisible to the text of the source code.

For example, suppose you'll try to answer a simple question about where Guava is used or where a particular logging library is used. In that case, you’ll find that while uses can occur in the code, the code-as-text may not have a reference to the library you are looking for. Imagine a logger instance inherited as a protected field from a base class defined in a binary dependency. The import statement that identifies which logging library that logger is coming from is IN the binary dependency, not in the text of the call site. A human would do no better in this situation.

This talk addresses improving AI accuracy for large-scale code refactoring by enhancing the data source. We’ll explore a state-of-the-art code data model called the Lossless Semantic Tree (LST), which is part of the open-source OpenRewrite auto-refactoring project. We’re finding that the LST and recipes are amazingly easy tools to equip LLMs with the data they need to make accurate decisions.

The common excuse for inaccuracy or incompleteness in output is that LLMs will improve, but I think the models are quite good enough right now. What they too often lack is the data to make inferences. We’ll show why when evaluating LLMs for large-scale automated refactoring:

* If it's based on text, you don't want it
* If it's based on AST, you don't want it

The State of AI in Large Scale Automated Refactoring

Global Java Community - Supported by Devnexus

Buscar {networkName}

Grupos

{PRO_NETWORK_NAME} está en Meetup Pro y tiene más de {PRO_MEMBER_COUNT} miembros en {PRO_GROUP_COUNT} grupos de Meetup en todo el mundo. Meetup Pro es la herramienta profesional para organizar redes de usuarios, socios, colaboradores y miembro, y comunicarse con ellos.

Global Java Community - Supported by Devnexus

Global Java Community - Supported by Devnexus

Quiénes somos

Buscar Global Java Community - Supported by Devnexus

Eventos previstos

Grupos