CARVIEW

MOTORHOMES

Select Language

HTTP/2 200 server: GitHub.com content-type: text/html; charset=utf-8 last-modified: Tue, 02 Sep 2025 02:42:36 GMT access-control-allow-origin: * strict-transport-security: max-age=31556952 etag: W/"68b6599c-9928" expires: Mon, 29 Dec 2025 04:43:43 GMT cache-control: max-age=600 content-encoding: gzip x-proxy-cache: MISS x-github-request-id: 9919:3157C7:8342E5:93AD86:695204A7 accept-ranges: bytes age: 0 date: Mon, 29 Dec 2025 04:33:43 GMT via: 1.1 varnish x-served-by: cache-bom-vanm7210023-BOM x-cache: MISS x-cache-hits: 0 x-timer: S1766982823.058255,VS0,VE217 vary: Accept-Encoding x-fastly-request-id: 8bc3e8d1e147990e416629eeb13bf868fe5c20ab content-length: 7271 Home | Practical-DL Workshop at IJCAI 2025

4th Workshop on Practical Deep Learning (Practical-DL 2025):
Toward Robust Compressed Foundation Models in the Real World

Workshop at IJCAI 2025

Overview

Compressed foundation models, particularly compressed large language models (LLMs), are increasingly deployed in real-world applications due to their efficiency advantages. Techniques such as quantization, pruning, and distillation significantly reduce memory consumption and latency, making it possible to run these models on resource-constrained devices, such as edge computing platforms and mobile applications. However, compression can also introduce critical robustness and security vulnerabilities, affecting both model reliability and trustworthiness. Compressed models are often more susceptible to adversarial attacks, hallucinations, jailbreak exploits, and degradation in out-of-distribution generalization, raising concerns about their safe deployment in high-stakes applications such as healthcare, finance, and content moderation.

This workshop places a particular emphasis on the robustness challenges of widely deployed compressed foundation models. While existing workshops may discuss general robustness issues in foundation models or focus on efficiency-driven compression, our workshop uniquely targets the intersection of model compression and robustness vulnerabilities. This perspective is crucial because compressed models are now widely used in practice, yet their resilience under real-world threats remains poorly understood and insufficiently addressed. For instance, quantized LLMs are more vulnerable to adversarial prompts, while pruned models may suffer from increased susceptibility to factual errors and hallucinations. Addressing these challenges is not just a theoretical exercise but a pressing need for ensuring secure and reliable deployment.

Call for Papers

Compressed large language models (LLMs) are increasingly adopted in practice due to their efficiency advantages. However, these models pose new challenges in robustness and security, raising concerns for their reliable deployment in real-world scenarios. We welcome research contributions related to the following (but not limited to) topics:

Robustness of Compressed Foundation Models
Strategies for Jointly Enhancing Compression Efficiency and Robustness
Compression for Robust Foundation Models
Efficient and Robust Architectural Designs
Robust and Efficient Inference Strategies
Privacy and Ethical Considerations in Compressed Models
Real-World Case Studies on Compressed Model Security and Privacy

Submission Format: Submissions papers (.pdf format) must use the IJCAI Article Template and be anonymized and follow IJCAI 2025 author instructions. The workshop considers two types of submissions: (1) Long Paper [7 pages]; (2) Extended Abstract [4 pages], including figures, tables and references.
Important: Our workshop will feature a Best Paper Award, and certificates will be presented during the workshop.
Submission Site: https://openreview.net/group?id=ijcai.org/IJCAI/2025/Workshop/Practical-DL
Submission Due: 30th May, 2025 GMT
Date and Location: August 29, as part of the IJCAI 2025 Satellite Event in Guangzhou, China.

Workshop Schedule

Schedule

Room: Ballroom FG

Keynote Speakers

Aaron Gilad Kusne

NIST

Yawei Li

ETH Zürich

Jie Zhang

Center for Frontier
AI Research
(CFAR), A*STAR

Haidong Kang

Northeastern University

Organizer

Xingyu Zheng

Beihang University

Haotong Qin

ETH Zürich

Aishan Liu

Beihang University

Jie Zhang

Center for Frontier
AI Research
(CFAR), A*STAR

Jiakai Wang

Zhongguancun Laboratory

Yulun Zhang

Shanghai Jiao Tong University

Olivera Kotevska

Oak Ridge National Laboratory

Xianglong Liu

Beihang University

Michele Magno

ETH Zürich

Dacheng Tao

Nanyang Technological University

Accepted Papers

Accepted Long Paper

Detection of Deepfake Videos and Audios on Social Media Platforms [Paper]
Charles Fross (Sam Houston State University); Qingzhong (Frank) Liu (Sam Houston State University)
Scalability and Noise Resilience in Q-Transformer [Paper]
Jun Wang (Hangzhou Dianzi University)
ASA: Adaptive Subquadratic Attention with Dynamic Complexity Adjustment [Paper]
Xinyuan Liu (Northwestern University)
Dialectical Chain Distillation: Transferring Dialectical Reasoning from Teacher–Student Interactions to Small Language Models [Paper]
Hairong Chen (Beijing Jiaotong University); Ruitao Li (Beijing Jiaotong University); Yuzhe Guo (Beijing Jiaotong University); Renshuai Tao (Beijing Jiaotong University); Huihui Bai (Beijing Jiaotong University); Yao Zhao (Beijing Jiaotong University) — Best Paper Award
Leveraging LLM-based Sentiment Analysis for Portfolio Optimization with Proximal Policy Optimization [Paper]
Kemal Kirtac (University College London); Guido Germano (University College London) — Outstanding Paper Award

Accepted Short Paper

Towards Leaner Architectures in Multimodal Language Models [Paper]
Prasanth (Affiliation TBD)
The Fault in Our LLM Leaderboards [Paper]
Pegah Jandaghi (University of Southern California); Kian Ahrabian (University of Southern California)