Results 51 to 60 of about 14,848 (289)

Self-Distillation for Randomized Neural Networks

IEEE Transactions on Neural Networks and Learning Systems
Knowledge distillation (KD) is a conventional method in the field of deep learning that enables the transfer of dark knowledge from a teacher model to a student model, consequently improving the performance of the student model. In randomized neural networks, due to the simple topology of network architecture and the insignificant relationship between ...
Minghui Hu, Ruobin Gao, Ponnuthurai Nagaratnam Suganthan   +2 more
openaire   +4 more sources

Large‐scale bidirectional arrayed genetic screens identify OXR1 and EMC4 as modifiers of αSynuclein aggregation

FEBS Open Bio, EarlyView.
Activation of the mitochondrial protein OXR1 increases pSyn129 αSynuclein aggregation by lowering ATP levels and altering mitochondrial membrane potential, particularly in response to MSA‐derived fibrils. In contrast, ablation of the ER protein EMC4 enhances autophagic flux and lysosomal clearance, broadly reducing α‐synuclein aggregates.
Sandesh Neupane, Lea Nikolić, Lorenzo Maraio, Thomas Goiran, Nathan Karpilovsky, Stefano Sellitto, Vangelis Bouris, Jiang‐An Yin, Ronald Melki, Edward A Fon, Adriano Aguzzi, Elena De Cecco   +11 more
wiley   +1 more source

Contrastive Learning or Masked Autoencoder? Understanding and Improving Self-Supervised Knowledge Distillation

IEEE Access
Lying at the intersection of self-supervised learning (SSL) and knowledge distillation (KD), Self-supervised KD (SSKD) differs from classical KD frameworks by assuming the teacher model is pretrained without labels.
Taegoo Kang, Sung-Ho Bae, Chaoning Zhang
doaj   +1 more source

Lower bounds for kernelizations [PDF]

, 2008
"Vegeu el resum a l'inici del document del fitxer adjunt"
Chen, Yijia, Flum, Jörg, Müller, Moritz, Universitat Autònoma de Barcelona. Centre de Recerca Matemàtica   +3 more
core  

Tailored Hierarchical Porous Copper Architectures via Three Dimensional Printing and Pressure‐less Sintering for Next‐Generation Lithium‐Metal Batteries

Advanced Engineering Materials, EarlyView.
A hierarchical porous copper current collector is fabricated via three‐dimensional printing combined with pressureless sintering to stabilize lithium metal anodes. The interconnected architecture lowers local current density, guides uniform Li deposition within pores, and suppresses dendrite growth.
Alok Kumar Mishra, Mukul Shukla
wiley   +1 more source

Adaptive Similarity Bootstrapping for Self-Distillation

, 2023
Most self-supervised methods for representation learning leverage a cross-view consistency objective i.e. they maximize the representation similarity of a given image's augmented views.
Bozorgtabar, Behzad, Lebailly, Tim, Stegmüller, Thomas, Thiran, Jean-Philippe, Tuytelaars, Tinne   +4 more
core  

Intermolecular Interactions as Driving Force of Increasing Multiphoton Absorption in a Perylene Diimide‐Based Coordination Polymer

Advanced Functional Materials, EarlyView.
This study uncovers the unexplored role of intermolecular interactions in multiphoton absorption in coordination polymers. By analyzing [Zn2tpda(DMA)2(DMF)0.3], it shows how the electronic coupling of the chromophores and confinement in the MOF enhance two‐and three‐photon absorption.
Simon Nicolas Deger, Anna Mauri, Yang Cui, Sebastian Josef Weishäupl, Ali Deniz Özdemir, Hamad Syed, Aleksandr Ovsianikov, Wolfgang Wenzel, Alexander Pöthig, Jürgen Hauer, Mariana Kozlowska, Roland Augustinus Fischer   +11 more
wiley   +1 more source

Graph Anomaly Detection Algorithm Based on Multi-View Heterogeneity Resistant Network

Information
Graph anomaly detection (GAD) aims to identify nodes or edges that deviate from normal patterns. However, the presence of heterophilic edges in graphs leads to feature over-smoothing issues. To overcome this limitation, this paper proposes the multi-view
Yangrui Fan, Caixia Cui, Zhiqiang Wang, Hui Qi, Zhen Tian   +4 more
doaj   +1 more source

Unleashing the Power of Machine Learning in Nanomedicine Formulation Development

Advanced Functional Materials, EarlyView.
A random forest machine learning model is able to make predictions on nanoparticle attributes of different nanomedicines (i.e. lipid nanoparticles, liposomes, or PLGA nanoparticles) based on microfluidic formulation parameters. Machine learning models are based on a database of nanoparticle formulations, and models are able to generate unique solutions
Thomas L. Moore, Cristiano Pesce, Antonietta Greco, Claudia Pisante, Greta Avancini, Valentina Di Francesco, Yosi Shamay, Paolo Decuzzi   +7 more
wiley   +1 more source

Clean, performance‐robust, and performance‐sensitive historical information based adversarial self‐distillation

IET Computer Vision
Adversarial training suffers from poor effectiveness due to the challenging optimisation of loss with hard labels. To address this issue, adversarial distillation has emerged as a potential solution, encouraging target models to mimic the output of the ...
Shuyi Li, Hongchao Hu, Shumin Huo, Hao Liang   +3 more
doaj   +1 more source