لطفا منتظر بمانید ...
0% Complete
صفحه اصلی
/
سی و چهارمین کنفرانس بین المللی مهندسی برق
SPNP: A Calibration-Aware and Sparse Output Normalization for Deep Neural Networks
نویسندگان :
Morteza Taheri
1
1- دانشگاه زنجان
کلمات کلیدی :
SPNP،softmax replacement،probability calibration
چکیده :
In this paper, we introduce Shifted Power Normalized Probabilities (SPNP), a simple closed-form output activation function designed as an alternative to Softmax for probability normalization. The proposed method performs sample-wise standardization of logits, followed by a minimal shift and power-based normalization. SPNP improves probability calibration and enables controllable sparsity without modifying network architectures or the underlying loss formulation. Experiments on CIFAR-100, CIFAR-10, and MNIST, where all models are trained using the same standard cross-entropy loss applied to logits, demonstrate that SPNP significantly improves calibration metrics such as ECE, MCE, and the Brier score, while maintaining or improving classification accuracy. In addition, SPNP produces sparse output distributions and introduces negligible inference-time computational overhead, making it a practical and reliable alternative to Softmax for calibrated probabilistic inference.
لیست مقالات
لیست مقالات بایگانی شده
The most descriptive surprise definition for brain’s EEG response to visual and auditory oddball tasks
Mohammad Mahdi Kiani - Zahra Mousavi - Hamid Aghajan
Improved Low Voltage Ride Through by A STATCOM Based on Neutral Point Piloted (NPP) Multilevel Inverter
Yousef Neyshabouri - Mohammad Farhadi-kangarlu
Probabilistic Dynamic Economic Dispatch in Presence of Wind Farms
Homayoun Berahmandpour - Shahram Montasar Kuhsari - Hassan Rastegar
Interferometric Direction Finding Based on Time-Modulated Array
Amirmahdi Javadi - Mohammadamin Latifi - Fereidoon Behnia
حسگر غیرتهاجمی تشخیص قندخون با استفاده از تکنیک مایکروویو بر مبنای تشدید فرکانسی
نازنین افشاری - سید محمد هاشمی - فاطمه گران قراخیلی
SWOT Analysis of the Mega Constellation Technology and Satellite Internet
Mohammad Bod - Parvin Sojoodi - Leila Mohammadi
Transfer learning using deep convolutional neural network for predicting dementia severity
Vahid Asayesh - Mehdi Dehghani - Majid Torabi Nikjeh - Sepideh Akhtari khosrowshahi
مدل هوشمند E-DEFS: چارچوبی پویا برای انتخاب ویژگی و پیشبینی عملکرد تحصیلی دانشآموزان
ساناز محمدی گرفمی - راهبه مجتهدی صفاری - ساسان حسینعلی زاده
Analyzing Large-scale PV Plant Controllers by Technical Performance Indices using MCS Method
Hooman Nasrazadani - Alireza Sedighi - Hossein Seifi
Using GA and ML to Improve LoRa Network Performance
Yas Hosseini Tehrani - Seyed mojtaba Atarodi
بیشتر
ثمین همایش، سامانه مدیریت کنفرانس ها و جشنواره ها - نگارش 44.7.2
