arXiv:2505. 24852v3 Announce Type: replace-cross Abstract: On-device learning at the edge enables low-latency, private personalization with improved long-term robustness and reduced maintenance costs.
By Douwe den Blanken, Charlotte Frenkel
The paper introduces a method for improving fine-grained, temporally aligned outputs in Speech Large Language Models (SpeechLLMs) by replacing absolute timestamps with relative timestamps, which reduces vocabulary size and enhances generalization. It proposes a hybrid fine‑tuning strategy that fully fine‑tunes the timestamp‑augmented embedding layer and language model head while applying LoRA to decoder layers, and introduces a masked timestamp training objective to prevent over‑reliance on ground‑truth timestamps. Experiments show significant gains in timestamp prediction accuracy without compromising transcription quality.
By Quanwei Tang, Zhiyu Tang, Xu Li, Dong Zhang, Shoushan, Guodong Zhou
arXiv:2608. 07524v1 Announce Type: new Abstract: Training deep learning models on variable long sequences poses significant computational challenges.
By Geng Zhang, Xuanlei Zhao, Kai Wang, Yang You
arXiv:2602. 16220v2 Announce Type: replace Abstract: Modeling multiscale patterns is crucial for long-term time series forecasting (TSF).
By Xu Zhang, Qitong Wang, Peng Wang, Wei Wang
arXiv:2607. 11746v1 Announce Type: new Abstract: With deep neural networks (DNNs) increasingly deployed on edge devices, hardware (HW)-aware optimization techniques--such as HW-aware compression and HW-aware neural architecture search (HW-NAS)--have become essential.
By Shambhavi Balamuthu Sampath, Behzad Shomali, Nael Fasfous, Moritz Thoma, Judeson Anthony Fernando, Lukas Frickenstein, Pierpaolo Mori, Manoj Rohit Vemparala, Alexander Frickenstein, Walter Stechele
The paper introduces the Variability-Aware Recursive Neural Network (VARNN), a residual-aware architecture for supervised time-series regression that learns a nonlinear, vector-valued residual representation from recent prediction errors. VARNN conditions subsequent predictions on this learned residual-memory state, mapping scalar prediction innovations into a short-context representation. Experiments on nine datasets across energy, healthcare, and environmental domains show that VARNN achieves lower test MSE than static, lag-based, and sequence-model baselines, and ablations confirm that the learned residual memory improves predictive accuracy over direct scalar residual feedback.
By Haroon Gharwi, Yue Dai, Kai Shu