ICFP/SPLASH 2025
Sun 12 - Sat 18 October 2025 Singapore
Wed 15 Oct 2025 13:55 - 14:10 at Orchid Small - LLMs for Code Generation Chair(s): Guannan Wei

This position paper proposes a fundamental shift in design- ing code generation models: treating reasoning depth as a controllable resource. Rather than being an incidental by- product of prompting, we argue that the trade-off between rapid, direct answers (“fast thinking”) and elaborate, chain- of-thought deliberation (“slow thinking”) must be explicitly managed. We contend that optimizing reasoning budgets across the entire model lifecycle—from synthetic data cre- ation and benchmarking to real-world deployment—can un- lock superior trade-offs among accuracy, latency, and cost. This paper outlines how adaptive control over reasoning can enrich supervision signals, motivate new multi-dimensional benchmarks, and inform cost-aware, security-conscious de- ployment policies. By viewing fast and slow thinking as complementary modes to be scheduled, we envision coding agents that think deep when necessary and act fast when possible.

Wed 15 Oct

Displayed time zone: Perth change

13:40 - 15:20
LLMs for Code GenerationLMPL at Orchid Small
Chair(s): Guannan Wei Tufts University
13:40
15m
Talk
W2GPU: Toward WebAssembly-to-WebGPU Program Translation via Small Language Models
LMPL
Mehmet Oguz Derin Unaffiliated
Media Attached
13:55
15m
Talk
Reasoning as a Resource: Optimizing Fast and Slow Thinking in Code Generation Modelsrecorded
LMPL
Zongjie Li The Hong Kong University of Science and Technology, Shuai Wang Hong Kong University of Science and Technology
14:10
15m
Talk
Ranking Formal Specifications using LLMsremote
LMPL
Deyuan (Mike) He Princeton University, Zhendong Ang National University of Singapore, Ankush Desai Amazon Web Services, Aarti Gupta Princeton University
14:25
15m
Talk
Challenges in C++ to Rust Translation with Large Language Models: A Preliminary Empirical Study
LMPL
Yanyan Yan Nanjing University, Yang Feng Nanjing University, Qi He Nanjing University, Jun Zeng Chongqing University, Baowen Xu Nanjing University
14:40
15m
Talk
The Modular Imperative: Rethinking LLMs for Maintainable Software
LMPL
Anastasiya Kravchuk-Kirilyuk Harvard University, Fernanda Graciolli Midspiral, Nada Amin Harvard University
14:55
15m
Talk
Programming Language Techniques for Bridging LLM Code Generation Semantic Gapsremote
LMPL
Yalong Du Harbin Institute of Technology, Shenzhen, Chaozheng Wang The Chinese University of Hong Kong, Huaijin Wang Ohio State University