paper-with-me

Papers

Parent-Guided Adaptive Reliability (PGAR): A Behavioural Meta-Learning Framework for Stable and Trustworthy AI

2026-01-07 · Anshum Rankawat arxiv

Parent-Guided Adaptive Reliability (PGAR) is a lightweight behavioural meta-learning framework that adds a supervisory "parent" layer on top of a standard learner to improve stability, calibration, and recovery under disturbances. PGAR computes three reflex-level signals (incident detection, overconfidence correction, and recovery memory) and fuses them into a bounded reliability index in [0,1]. This index continuously modulates the learner's effective learning rate, reducing update magnitude during instability and restoring it as reliability improves. We provide a Lyapunov-based proof sketch establishing bounded adaptation of the reliability dynamics under mild assumptions (smooth loss, descent direction, and bounded reflex outputs). Empirical evaluations on representative learning tasks show improved calibration, reduced loss variance, and faster recovery compared to standard optimizers, while retaining computational simplicity. PGAR functions as a plug-in reliability layer for existing optimization and learning pipelines, supporting interpretable reliability traces in safety-relevant settings.

📄 PDF Abstract BibTeX arXiv:2601.06167

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

DeepGAR: Deep Graph Learning for Analogical Reasoning

2022-11-19 · Chen Ling, Tanmoy Chowdhury, Junji Jiang, Junxiang Wang 외

Analogical reasoning is the process of discovering and mapping correspondences from a target subject to a base subject. As the most well-known computational method of analogical reasoning, Structure-Mapping Theory (SMT) …

Graph Learning

Group Activity Recognition using Unreliable Tracked Pose

2024-01-06 · Haritha Thilakarathne, Aiden Nibali, Zhen He, Stuart Morgan

Group activity recognition in video is a complex task due to the need for a model to recognise the actions of all individuals in the video and their complex interactions. Recent studies propose that optimal performance i…

Activity RecognitionAllGroup Activity RecognitionOptical Flow Estimation

The human behavioural immune system is a product of cultural evolution

2020-08-30 · Edwin S. Dalmaijer, Thomas Armstrong

Disgust is a basic emotion that serves to avoid contaminants, and is central to the behavioural immune system. While disgust-motivated avoidance occurs in bonobos and chimpanzees, humans show uniquely high levels of cont…

Nutrition

A Behavioural Foundation for Natural Computing and a Programmability Test

2013-03-23 · Hector Zenil

What does it mean to claim that a physical or natural system computes? One answer, endorsed here, is that computing is about programming a system to behave in different ways. This paper offers an account of what it means…

Rethinking Transparent Object Grasping: Depth Completion with Monocular Depth Estimation and Instance Mask

2025-08-04 · Yaofeng Cheng, Xinkai Gao, Sen Zhang, Chao Zeng 외 arxiv

Due to the optical properties, transparent objects often lead depth cameras to generate incomplete or invalid depth data, which in turn reduces the accuracy and reliability of robotic grasping. Existing approaches typica…

Monocular Depth EstimationDepth CompletionRobotic Grasping