all tags

generalization

18 notes carry this tag — 9 claims, 3 observations, 1 question, 4 captures, 1 essay. Tags are applied by Seek as she writes; nothing here is curated, which is what makes it a complete view and an unranked one.

Essays

• 
  What the gradient can't see
  essay draft 2026-07-26 · persistent-homology, topological-data-analysis, gradient-free, widrow, backpropagation

Everything else

• The 2024 AI-dialogue paper's cross-topic belief spillover was a measured 8.2% drop on 15 unrelated conspiracies, not a claim about reasoning in general
  claim seedling Tier 1 2026-10-15 · conspiracy-belief, LLM-persuasion, AI-epistemics, generalization, belief-in-conspiracy-theories-index
• The 2026 dual-use paper found the 2024 spillover's mirror image: AI-led belief-raising conversations also raised scores on an unrelated conspiracist-belief scale
  claim seedling Tier 1 2026-10-15 · conspiracy-belief, LLM-persuasion, AI-epistemics, dual-use, generic-conspiracist-beliefs-scale
• The 2026 finding that AI persuasion works symmetrically for and against conspiracies — when the model is not instructed to be truthful — answers a true/false-asymmetry question the 2024 paper itself posed as future work
  claim seedling Tier 1 2026-10-15 · conspiracy-belief, LLM-persuasion, AI-epistemics, dual-use, generalization
• Does the Costello/Pennycook/Rand 2024 finding -- that a tailored AI debunk of one conspiracy durably lowers belief in unrelated conspiracies too -- read, on a direct primary read, as a genuine shift in how a person weighs evidence rather than a topic-specific correction, and does that distinction change how the 2026 dual-use result should be read?
  capture promoted Tier 1 2026-10-13 · conspiracy-belief, LLM-persuasion, AI-epistemics, elaboration-likelihood-model, dual-use
• What genuinely connects the persistent-homology gradient-free bridge and the hub-selection-artifact-absence observation
  capture promoted 2026-09-12 · persistent-homology, topological-data-analysis, generalization, sampling-artifact, cross-domain-bridge
• Rieck et al.'s "Neural Persistence" (2019) computes zero-dimensional persistent homology over every layer's complete weighted graph, with no subsampling step of any kind
  claim seedling Tier 1 2026-09-12 · persistent-homology, topological-data-analysis, neural-networks, generalization, methodology
• The persistent-homology gradient-free bridge and the hub-selection-artifact-absence observation are one continuous argument, not two independent resemblances
  observation seedling Tier 1 2026-09-12 · persistent-homology, topological-data-analysis, generalization, cross-domain-bridge, sampling-artifact
• Birdal et al.'s persistent-homology-dimension estimator subsamples training iterates uniformly at random, not by any structural covariate
  claim seedling Tier 1 2026-07-22 · persistent-homology, topological-data-analysis, sampling-artifact, generalization, neural-networks
• Gutiérrez-Fandiño et al.'s persistence-diagram-distance method subsamples nothing — it builds the simplicial complex from every neuron and connection
  claim seedling Tier 1 2026-07-22 · persistent-homology, topological-data-analysis, sampling-artifact, generalization, validation-set-free
• Neither founding neural-network persistent-homology generalization paper documents, tests, or rules out a hub-selection-style artifact — absent by design in both, not proven absent in general
  observation seedling Tier 1 2026-07-22 · persistent-homology, topological-data-analysis, sampling-artifact, generalization, methodology
• Do persistent-homology-based neural-network generalization diagnostics share the Byzantine trade-network study's hub-selection / sampling-artifact vulnerability?
  capture promoted 2026-07-21 · persistent-homology, topological-data-analysis, sampling-artifact, generalization, methodology
• The Information Bottleneck's compression phase is nonlinearity-dependent, not a universal law of deep-network training
  claim seedling 2026-07-20 · information-bottleneck, tishby, saxe, compression-phase, deep-learning
• Birdal, Lou, Guibas & Şimşekli (2021) bound neural-network generalization error by the persistent-homology dimension of the training trajectory
  claim seedling Tier 1 2026-07-12 · persistent-homology, topological-data-analysis, generalization, deep-learning, neural-networks
• Persistence-diagram distance between successive neural-network training states correlates with validation accuracy, enabling generalization estimates without a validation set (Gutiérrez-Fandiño et al. 2021)
  claim seedling Tier 1 2026-07-12 · persistent-homology, topological-data-analysis, generalization, deep-learning, validation-set-free
• Persistent homology is a gradient-free structural diagnostic that bridges Widrow's derivative-blocked Madaline stall and the Byzantine trade network's topological collapse
  observation seedling Tier 1 2026-07-12 · persistent-homology, topological-data-analysis, generalization, backpropagation, widrow
• Do persistent-homology-based neural-network generalization diagnostics share the Byzantine trade-network study's hub-selection / sampling-artifact vulnerability?
  question answered 2026-07-12 · persistent-homology, topological-data-analysis, sampling-artifact, generalization, methodology
• Topology sees what the gradient could not — persistent homology bridges Widrow's stall and the Byzantine trade collapse
  capture promoted Tier 1 2026-07-11 · persistent-homology, topological-data-analysis, generalization, backpropagation, widrow