Catastrophic Interference in Connectionist Networks¶
McCloskey, M., & Cohen, N. J. (1989). Catastrophic Interference in Connectionist Networks: The Sequential Learning Problem. Psychology of Learning and Motivation.
Cited by¶
3 citations across 3 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Reconsolidation
- In ML fine-tuning, loading a checkpoint, applying gradients, and saving updated weights reconsolidates prior learning under new evidence, and catastrophic forgetting is reconsolidation-driven loss.
This sourceIdentifies catastrophic forgetting, where training on new data overwrites prior learning — the reconsolidation pathology in neural networks.
- In ML fine-tuning, loading a checkpoint, applying gradients, and saving updated weights reconsolidates prior learning under new evidence, and catastrophic forgetting is reconsolidation-driven loss.
Mechanisms¶
- Competing-Case Spacing Rule
- This is the everyday face of catastrophic interference, the tendency, first sharply characterized in connectionist networks, for new learning that overlaps old learning to wipe it out when the two are packed together.
This sourceCharacterizes catastrophic interference in connectionist networks, where new overlapping learning can sharply erase previously learned information.
- This is the everyday face of catastrophic interference, the tendency, first sharply characterized in connectionist networks, for new learning that overlaps old learning to wipe it out when the two are packed together.
- Map Difference and Integrity Check
- Its strength is that it turns "the map feels off" into a located, evidenced divergence, and it is the sensor for catastrophic interference
This sourceShows that sequential new learning can alter weights used for older representations, producing catastrophic interference.
- Its strength is that it turns "the map feels off" into a located, evidenced divergence, and it is the sensor for catastrophic interference
Verification¶
Does it exist? Confirmed. This work's DOI resolves to a registered record, which fixes its identity. That is all it fixes.
Does it back the claim? Not recorded. Neither this nor any other of the 3 citations of this work carries a recorded support check.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Registry ID ref:225f050107c8 · see in the full table