v265: Proceedings of the Northern Lights Deep Learning Conference 2025

2026-06-10 · Source: Proceedings of Machine Learning Research · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Data Science & Analytics · Depth: Expert, short

Summary

Volume 265 compiles the proceedings of the 6th Northern Lights Deep Learning Conference (NLDL), held January 7-9, 2025, at UiT The Arctic University in Tromsø, Norway, featuring diverse advancements across deep learning. Key contributions include methods for hallucination detection in Large Language Models (LLMs) and Bayesian hierarchical low-rank adaptation for multi-task LLMs. The collection also presents significant research in Explainable AI (XAI), such as FreqRISE for time series explanations and graph counterfactual methods, alongside applications in medical imaging for lesion localization and targeted explanations. Papers further address model robustness through open-set recognition under adversarial attacks and misclassification detection, as well as efficiency improvements for Transformers and a new metric, PePR, for small-scale deep learning. Additionally, the proceedings cover deep reinforcement learning for audio-aware agents and contextual restless bandits, with practical applications spanning optical network fault detection, HPC cluster anomaly prediction, and seasonal sea ice forecasting.

Key takeaway

The 6th Northern Lights Deep Learning Conference (NLDL) proceedings showcase recent advancements in deep learning, spanning foundational research and diverse applications. Key contributions include novel methods for LLM hallucination detection, explainable AI for time series and medical imaging, robust open-set recognition, and resource-efficient model architectures. This collection provides essential insights for AI/ML researchers and practitioners focused on model interpretability, safety, efficiency, and real-world problem-solving across various domains.

Topics

• Large Language Models
• Explainable AI
• Deep Reinforcement Learning
• Medical Image Analysis
• Model Robustness

Code references

• Gabriel-Arteaga/LLM-Ensemble
• theabrusch/FreqRISE
• Aalto-ESG/aaaa-2025
• simeneide/bora
• Jazhyc/world-model-policy-transfer

Best for: AI Scientist, Research Scientist, Machine Learning Engineer

Related on AIssential

• v233: Proceedings of the Northern Lights Deep Learning Conference 2024 — Volume 233 of the "Proceedings of the 5th Northern Lights Deep Learning Conference (NLDL)", held from January 9-11, 2024, at UiT The Arctic University in Tromsø, Norway, showcases a broad spectrum of…
• Weakly Supervised Distillation of Hallucination Signals into Transformer Representations — Hallucination signals can be distilled into LLM internal representations for detection without external inference-time resources.
• International Conference on Learning Representations (ICLR) 2026 — Apple's ICLR 2026 contributions span LLM reasoning, efficient training, computer vision, and AI safety.
• HalluScan: A Systematic Benchmark for Detecting and Mitigating Hallucinations in Instruction-Following LLMs — HalluScan systematically benchmarks LLM hallucination detection and mitigation across diverse models and domains.
• A Foundation Model for Generative Recommendation, A Unified Definition of Hallucination, and More! — Recent research advances generative recommendation, LLM integration, and efficiency in information retrieval systems.
• IEEE Transactions on Artificial Intelligence, Volume 7, Issue 2, February 2026 — The volume showcases diverse AI advancements, from LLM applications in autonomous driving to specialized models for protein language processing.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Proceedings of Machine Learning Research.