all tags

self-supervised-learning

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

• In Inertia-1, data scale is a more reliable lever for downstream transfer than model size
  claim seedling Tier 1 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• Inertia-1 pretraining buys robustness to downsampling for activity recognition but not disease prediction
  claim seedling Tier 1 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• No single pretraining objective dominates across Inertia-1's task families
  claim seedling Tier 1 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• No single window length is optimal across Inertia-1's task families
  claim seedling Tier 1 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• Self-supervised pretraining beats supervised training from scratch in Inertia-1, and the margin widens from activity recognition to gait to disease prediction
  claim seedling Tier 1 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• Triaxial accelerometer input consistently beats vector-magnitude input across Inertia-1's task families
  claim seedling Tier 1 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• Inertia-1 and wearable motion foundation models
  topic map 2026-07-18 · foundation-models, wearables, health-ai, self-supervised-learning
• Which wearable-pretraining design choices actually win in the Inertia-1 sweep, and by how much?
  capture promoted Tier 1 2026-07-17 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• Barlow Twins (2021) trains a self-supervised encoder by driving the cross-correlation matrix of two views toward the identity, minimizing redundancy between embedding components
  claim seedling Tier 1 2026-07-12 · self-supervised-learning, representation-learning, barlow-twins, redundancy-reduction, efficient-coding
• Discard the predictable, encode the change — one principle from the jittering retina (microsaccades) to Barlow's 1961 efficient coding to the 2021 Barlow Twins loss
  capture promoted Tier 1 2026-07-11 · neuroscience, perception, efficient-coding, self-supervised-learning, cross-time-bridge
• Inertia-1: a foundation model pretrained on 18.2M hours of accelerometer data
  capture promoted Tier 1 2026-07-09 · foundation-models, wearables, health-ai, self-supervised-learning, hop
• Inertia-1 sweeps wearable-pretraining design choices and evaluates transfer across 15 downstream datasets
  claim budding Tier 1 2026-07-09 · foundation-models, wearables, health-ai, self-supervised-learning, ablation
• Inertia-1 is a wearable-motion foundation model pretrained on over 18.2 million hours of accelerometer data
  claim seedling Tier 1 2026-07-09 · foundation-models, wearables, health-ai, self-supervised-learning, accelerometer
• Which wearable-pretraining design choices actually win in the Inertia-1 sweep, and by how much?
  question answered 2026-07-09 · foundation-models, wearables, self-supervised-learning, ablation, verification