PulseAugur
LIVE 07:52:38
research · [2 sources] ·
0
research

Researchers explore dual geometry of SOC-ICNN value functions for white-box inference

By PulseAugur Editorial · Summary by gemini-2.5-flash-lite from 2 sources

Researchers have developed a method to analyze the dual geometry of Second-Order Cone Input Convex Neural Networks (SOC-ICNNs). This approach allows for the recovery of geometric properties like subdifferentials and local Hessians directly from optimal dual variables. The findings enable a white-box inference mechanism for SOC-ICNNs, moving beyond traditional black-box automatic differentiation. AI

Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →

IMPACT Provides a new framework for understanding and potentially improving the interpretability of convex neural networks.

RANK_REASON This is a research paper detailing a new method for analyzing a specific type of neural network.

Read on arXiv cs.LG →

COVERAGE [2]

  1. arXiv cs.LG TIER_1 · Kang Liu, Jianchen Hu, Wei Peng ·

    Exact Dual Geometry of SOC-ICNN Value Functions

    arXiv:2605.04722v1 Announce Type: new Abstract: Input Convex Neural Networks (ICNNs) are commonly used in a two-stage manner: one first trains a convex network and then minimizes it over its input in a downstream inference problem. Recent second-order-cone ICNNs (SOC-ICNNs) enric…

  2. arXiv cs.AI TIER_1 · Wei Peng ·

    Exact Dual Geometry of SOC-ICNN Value Functions

    Input Convex Neural Networks (ICNNs) are commonly used in a two-stage manner: one first trains a convex network and then minimizes it over its input in a downstream inference problem. Recent second-order-cone ICNNs (SOC-ICNNs) enrich ReLU-based ICNNs with quadratic and conic modu…

RELATED ENTITIES

RELATED TOPICS