question

active

question:how-are-model-spaces-constructed-and-explored-in-the-absence-of-a-full-model-bottom-up-structure-learning

How are model spaces constructed and explored in the absence of a full model (bottom-up structure learning)?

Research gap explicitly identified: the paper uses top-down BMR from a full model, avoiding this challenge

Source paper

extracted_from

Active Inference, Curiosity and Insight

(2017) · Karl Friston · Marco Lin · Chris Frith · Giovanni Pezzulo +2

Neighborhood — ranked by edge-count

Concepts (1)

concept

Active Inference, Curiosity and Insight (Friston et al., 2017)
associated_with
The primary source paper being extracted

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Connectionist models can exhibit learning bottom-up, without centralised control or an external teacher, and without any performance feedback applied at the system level.claim0.786
Key property of distributed unsupervised learning.
Unsupervised learning builds a low-dimensional model of the input data.claim0.770
Clarifies what unsupervised learning does.
Bottom-up learning creates a non-decomposable whole (attractors that are non-linearly separable functions of the inputs and depend on the system’s own internal history), which means that credit assignment or reward at the level of individual parts becomes ineffective.claim0.753
Explains how collective cognition becomes irreducible to parts.
Base models are good modellers of worlds but not of their own state, because they lack a developed self-model initially.claim0.744
Observation about asymmetry in base model capabilities.
Connectionist models of cognition and learning identify conditions where collective intelligence can arise bottom-up, using only distributed learning mechanisms without system-level or global feedback.claim0.739
Central claim about the power of connectionism.
Bottom-up interpretability explains computation in the model's own terms rather than imposing top-down abstractionsclaim0.739
VPD is positioned as advancing a paradigm shift from top-down mechanistic interpretability (activation-based) to parameter-centric, data-driven discovery.
There are fewer representations competent for N tasks than M<N tasks, so training more general models should yield fewer possible solutionshypothesis0.737
Selective pressure toward convergence via task generality
When a model discovers that its outputs produce effects, it accelerates learning through in-context learning, analogous to lucid dreaming.claim0.735
Describes scaffolding method and the model's meta-learning loop.