A scientist with dual industry and academic affiliation shares insights on developing a fundamental scientific theory of machine learning, based on approximately seven years of work. The post outlines thoughts on achieving a true science of deep learning.
This article develops a unified account of Mediative Fuzzy Logic, from its type-1 foundations to type-2, type-3, and quantum extensions. It characterizes the mediative operator, models mediative truth values, and introduces a propositional system, establishing its soundness and paraconsistency.
This paper investigates the size complexity and decidability of first-order progression, a method for updating knowledge bases in AI. It demonstrates that for specific action classes within the Situation Calculus framework, first-order progression grows polynomially and maintains decidability when using certain knowledge base fragments.
This content outlines a highly abstract conceptual map or theoretical framework, possibly for an AI system, detailing concepts like symbolic and physical time, and an energetic triad. It describes a reversible topology for a 'Tightened Terminal Resonant Platform of Physicality,' with an explicit mention of Perplexity.ai.
This paper introduces a finite sheaf-theoretic framework to detect scientific theory shift in AI agents. It focuses on identifying when an existing representational framework needs extension due to local-to-global obstruction rather than simple deformation.
This work introduces a classification of learning dynamics (scalar-reducible vs. scalar-irreducible) and demonstrates that scalar-irreducible dynamics enable endogenous regime switching, crucial for autonomous intelligence. It proposes a new dynamical paradigm for regime exploration without external scheduling.
This work uses the conditional probabilities computed by LLMs to situate them within the mathematical theory of stochastic processes. It presents a model-agnostic probabilistic token attribution measure, using Bayes' rule to capture the model's internal representation of the distribution over token sequences.
This content introduces continuous time in Active Inference, highlighting that an agent's posterior belief about the world's state moves with its dynamics, integrating predictive coding and control theory. It explains how this involves gradient descent in generalized coordinates, encompassing position, velocity, and acceleration, for a unified system.
This content explores the concept of artifacts as a form of memory that extends beyond an agent's boundaries. It discusses how external objects can aid in cognition and information storage.
This article explores the concept of maintaining functional continuity in systems, potentially AI, without explicit reliance on memory structures. It may discuss philosophical implications or architectural designs achieving this state.