Pith. sign in

REVIEW 3 major objections 4 minor 59 references

Data augmentation as a framework for modeling hippocampal contributions to generalization

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper argues that hippocampal replay, preplay, and retrieval can be understood as offline and online forms of data augmentation, and that this framing lets theories of memory be tested as models running on the same sensory input animals

desk verdict A clear, honest perspective that offers a useful but unproven framing; worth engaging, provided reviewers push for specificity. read the letter →

arxiv 2608.01297 v1 pith:NEIBILJR submitted 2026-08-02 q-bio.NC

Tyler Bonnen , Andrew Kyle Lampinen This is my paper · ORCID
classification q-bio.NC
keywords hippocampusdataaugmentationgeneralizationmemoryreplayepisodicstimulus-computablemodelinglinkingfunctionssystemsconsolidation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes that data augmentation—the machine-learning practice of synthesizing new training examples from existing ones—offers a principled way to model how the hippocampus supports generalization. It identifies two timescales: offline augmentation, in which replay and consolidation re-factor stored experiences so cortical systems extract latent structure; and online augmentation, in which retrieved experiences are re-projected at test time to support flexible, zero-shot inference. The paper argues these operations correspond to functions attributed to the hippocampus, from spatial navigation to relational reasoning. Its central proposed payoff is methodological: if theories of hippocampal function are instantiated as stimulus-computable models operating on the same sensory input animals receive, then model design choices can serve as formal linking functions connecting experimental data to theoretical claims.

What carries the argument

The load-bearing machinery is the concept of data augmentation itself, defined as generating additional learning signals from an existing experience by transformation. The paper splits the concept into offline augmentation (fixed transformations applied during training to build robust representations) and online augmentation (flexible, context-dependent transformation of retrieved content at test time). The second key piece is the stimulus-computable model: a model that operates on the same sensory inputs as an experimental subject, making the model's operations a formal linking function between theory and data. Together they let claims about hippocampal replay, preplay, and retrieval be exp

What would settle it

A concrete test: build a stimulus-computable model with the proposed offline and online augmentation operations and train it on the exact sensory trajectories animals receive; the framework fails if the model cannot reproduce hippocampal-dependent behaviors (e.g., novel shortcut formation, transitive inference), or if a model with no augmentation of stored experiences matches behavior equally well. A complementary neural test: if replay always veridically recapitulates stored episodes and never re-factors them, the proposed equivalence is contradicted.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that hippocampal contributions to generalization can be characterized as data augmentation: a single stored experience is a single projection of an underlying structure, and the hippocampus preserves experience in a form that supports re-projection, recombination, and reversal. Offline, replayed experiences are interleaved with ongoing learning, allowing neocortical circuits to build general representations from sets of related events. Online, recalled experiences are re-factored in a task-dependent way to support inferences that were never trained. The paper does not present this as a wholly new theory of hippocampal function; rather, it presen

Load-bearing premise

The load-bearing premise, acknowledged in the paper's caveat that it is not proposing a new theory of the hippocampus (Section 3), is that the operations the hippocampus applies to stored experiences during replay, preplay, and retrieval are algorithmically equivalent to data augmentation; if that equivalence is only a metaphor, the proposed stimulus-computable models would not capture hippocampal function.

Editorial extensions

If this is right

  • Theories of replay that currently specify only which experiences are replayed can be extended to specify how those experiences are transformed, and each choice yields distinct behavioral predictions.
  • Long-standing disagreements—such as whether shortcut planning is generated within the hippocampus or through hippocampal-cortical exchange—become empirically testable modeling choices.
  • A single stimulus-computable framework can span navigation, transitive inference, and other hippocampal-dependent behaviors.
  • Verbal theories lose the ability to leave key operations implicit; model construction forces commitments about which operations are hippocampal and which emerge from interaction with other structures.
  • Empirical paradigms rich enough for naturalistic generalization become priority targets, since the framework's value depends on models confronting the same sensory data as animals.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper, this framing predicts that lesioning or silencing the hippocampus during a task should disproportionately impair zero-shot relational transformations (e.g., reversing a learned relation) while leaving well-learned direct associations intact.
  • Beyond the paper, the offline/online distinction suggests a graded spectrum of augmentation complexity—from fixed geometric transforms to model-based semantic elaboration—that could be indexed experimentally by how replay content changes with task demands.
  • Beyond the paper, the same logic applied to word-list or paired-associate paradigms would predict that amnesic patients' errors mirror the absence of specific augmentation operations (e.g., no reversal or composition), not just loss of memory strength.
  • Beyond the paper, if hippocampal replay truly is augmentation, then the statistics of replay content in naturalistic settings should match the augmentation policy that maximizes a stimulus-computable model's downstream generalization on the same task.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. Abstract, Sections 1–3, Boxes 1–4: The paper proposes data augmentation as a unifying framework for understanding hippocampal contributions to generalization. It distinguishes offline augmentation (transforming training data to build robust representations) from online augmentation (retrieving and re-factoring stored experiences at test time to support zero-shot inference). It maps these strategies onto hippocampal phenomena: offline replay during consolidation is likened to training-time augmentation, while online retrieval/replay in navigation and relational reasoning is likened to test-time augmentation. The authors argue that this framing can connect theories of hippocampal function to stimulus-computable models that operate on the same sensory inputs as experimental subjects, thereby providing 'linking functions' between neural/behavioral evidence and theoretical claims (Box 2, Box 4). They explicitly state that they are not proposing a new theory but a modeling framework, and they identify as a next step the construction of a stimulus-computable model for a well-characterized behavior (Section 3).

Significance. The paper's value lies in synthesis and methodological proposal. It brings together a broad range of ML and neuroscience findings and offers a simple taxonomy (offline/online) that could help organize otherwise disparate replay phenomena. The proposal to use stimulus-computable models as linking functions is a concrete and welcome step toward making verbal theories testable, and the authors are explicit about the challenges (Section 3). The paper is also honest in acknowledging that no such model is built, so the central promise is currently programmatic. If the mapping can be made precise and instantiated, the framework could provide a unifying computational vocabulary for CLS, predictive-map, and compositionality accounts. However, the paper's current contribution as a framework is conditional on that formalization.

major comments (3)
  1. [Section 2 / Box 4] The central analogy is asserted rather than derived. The paper defines data augmentation as requiring 'an experience, and operations to transform it' (Section 2), but never specifies what makes a transformation valid for a given task. In image augmentation, rotation is useful because object identity is invariant to rotation (Section 1.1); without a corresponding invariance structure, hippocampal reversal, composition, and coordinate re-projection are just relabeled transformations. As a result, any replay phenomenon—veridical, reverse, compositional, or off-trajectory—can be described as 'augmentation,' and the framework is unfalsifiable as stated. Box 4 claims distinct testable predictions, but those predictions require a formally specified augmentation set and validity conditions. Section 3 concedes that building such a model is 'a natural first step,' so the central promise is unteste
  2. [Section 1.2 and footnote 2] The online/offline distinction is stretched in a way that undermines specificity. Section 1.2 defines online augmentation as retrieving relevant information and providing it as context, and footnote 2 explicitly acknowledges that 'interpreting retrieval as online data augmentation is atypical.' Under this definition, any episodic memory system that retrieves experiences at test time performs online augmentation; the framework therefore does not distinguish hippocampal function from generic memory retrieval. The authors need to state which additional operations (e.g., reversal, composition, re-projection) must be applied to the retrieved content, and which behavioral signatures would fail to count as augmentation. Without this constraint, the claimed mapping to the hippocampus is trivial.
  3. [Box 4 / Section 3] Box 4 and Section 3 state that stimulus-computable models can serve as linking functions and that model design choices can be evaluated by fit to animal behavior. However, no such model is presented, and the navigation example in Box 4 does not specify the sensory input, model class, or evaluation metric. The promise of a formal linking function is therefore programmatic rather than demonstrated. For a perspective, this is acceptable, but the authors should more clearly distinguish between a research agenda and an established framework, and should provide at least one falsifiable prediction that could be tested before full stimulus-computable modeling is achieved.
minor comments (4)
  1. [Introduction] In the introductory paragraph, 'using the some empirical framework' should read 'using the same empirical framework.'
  2. [Section 2 / Box 1] There are typos: 'enablie' should be 'enable' in Section 2, and 'unifyingnormative' in Box 1 is missing a space.
  3. [Section 3] 'the identifying a novel shortcut' is ungrammatical; please rephrase.
  4. [References] Reference [28] is listed as 'Nature, under review'; please update to an archival citation or mark it clearly as a preprint. Several other references are future-dated preprints; consider adding version numbers or DOIs.

Circularity Check

1 steps flagged · score 2.0 of 10

Minor self-definitional breadth: hippocampal replay/retrieval is data augmentation by the paper's own broad definition, but the central modeling proposal is independent.

  1. self definitional [Section 2, 'The analogy between data augmentation and hippocampal function' (and footnote 2, Section 1.2)]
    "In essence, data augmentation requires two components: an experience, and operations to transform it. We suggest the hippocampus preserves information related to single events in a way that supports subsequent transformation... This is data augmentation in the precise sense that we began with: the hippocampus supplies an experience, via replay, which can be operated on by other neural structures."

    The paper defines data augmentation so broadly—'an experience, and operations to transform it'—that any refactoring of stored experience counts as augmentation. Hippocampal replay and retrieval are then described precisely as supplying experiences and transforming them, so the central mapping from hippocampal phenomena to data augmentation is true by definition rather than empirically derived. The paper is transparent that it is offering a perspective, not a novel theory, and it does not use this mapping to make quantitative predictions. Still, the applicability of the framework to the hippocampus is guaranteed by the breadth of the definition, which is a mild self-definitional circularity.

full rationale

The paper is a perspective/review, not a derivation. It does not fit parameters to data, make quantitative predictions, or invoke a uniqueness theorem. The central claim—that hippocampal replay, preplay, and retrieval can be viewed as data augmentation—is an analogy/definitional reframing. The paper explicitly acknowledges this in Section 2: 'This is not a novel claim about hippocampal function.' The only circular element is that 'data augmentation' is defined broadly enough (an experience plus operations to transform it) that hippocampal replay/retrieval trivially qualifies; the mapping is true by construction. However, the paper's actual proposal—building stimulus-computable models where augmentation strategies serve as linking functions—is independent and in principle falsifiable, though it remains unimplemented ('A natural first step is to build such a model'). Self-citations (e.g., Bonnen et al. 2021; Lampinen et al. 2026) are used as evidence of feasibility and prior modeling, not as the sole justification for the central claim. No load-bearing self-citation chain or fitted-input-called-prediction pattern is present. Overall circularity is minor, so the score is 2.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

This paper is a perspective, so the axiom ledger captures the key domain assumptions needed for the proposed framework to be valid. No free parameters or invented entities are introduced.

assumptions (3)
  • domain assumption The hippocampus preserves information related to single events in a way that supports subsequent transformation.
    Stated in Section 2, 'We suggest the hippocampus preserves information...'. This is a key premise for the analogy; if false, the mapping between stored experiences and data augmentation operations fails.
  • domain assumption Stimulus-computable models can be built for hippocampal memory tasks operating over the same sensory data as experimental subjects.
    Box 2 and Conclusions acknowledge this is not yet realized ('Realizing this promise is now potentially a feasible undertaking'). The claim that data augmentation can provide linking functions depends on this feasibility.
  • ad hoc to paper Data augmentation operations (e.g., reversal, composition, re-projection) are computationally equivalent to the transformations hippocampal replay/preplay applies to stored experiences.
    This is the central framing of the paper, assumed rather than derived. It is the bridge between ML augmentation and hippocampal function, and is not independently established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data augmentation as a framework for modeling hippocampal contributions to generalization." pith.science (2026). https://pith.science/paper/NEIBILJR

@misc{pith2026260801297,
  author       = {Pith},
  title        = {Pith review of: Data augmentation as a framework for modeling hippocampal contributions to generalization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NEIBILJR}},
  note         = {Machine review of arXiv:2608.01297}
}
read the original abstract

The hippocampus plays a critical role in generalization, enabling us to flexibly repurpose prior experiences to perform novel tasks. Here we suggest that data augmentation---a machine learning strategy to improve generalization by refactoring prior experience---offers a useful framework to conceptualize and model hippocampal function. We begin by outlining how data augmentation operates across two timescales: the traditional ``offline'' setting, where refactoring training data yields more general representations, and an ``online'' setting, where retrieved experiences can be flexibly refactored at test time to support zero-shot inference. We suggest that these `offline' and `online' computational strategies map onto functions supported by the hippocampus. Critically, we argue that these computational tools can be leveraged to develop formal `linking functions' between experimental evidence and theoretical claims, such that a unified modeling approach can be used to predict the diverse behaviors that depend on the hippocampus---from navigating in high-dimensional sensory environments to more abstract inferences. We hope this perspective, and the modeling strategies it makes available, will support new efforts to formalize and evaluate theories of hippocampal function.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 2 linked inside Pith

  1. [1]

    A cortical–hippocampal system for declarative memory.Nature Reviews Neuroscience, 1(1):41–50, 2000

    Howard Eichenbaum. A cortical–hippocampal system for declarative memory.Nature Reviews Neuroscience, 1(1):41–50, 2000

    2000

  2. [2]

    Medial temporal lobe amnesia: Gradual acquisition of factual information by nondeclarative memory.Journal of Neuroscience, 22(13):5741–5748, 2002

    Peter J Bayley and Larry R Squire. Medial temporal lobe amnesia: Gradual acquisition of factual information by nondeclarative memory.Journal of Neuroscience, 22(13):5741–5748, 2002

    2002

  3. [3]

    Hippocampal replay is not a simple function of experience.Neuron, 65(5):695–705, 2010

    Anoopum S Gupta, Matthijs AA Van Der Meer, David S Touretzky, and A David Redish. Hippocampal replay is not a simple function of experience.Neuron, 65(5):695–705, 2010

    2010

  4. [4]

    Hippocampal place cells construct reward related sequences through unexplored space.Elife, 4:e06063, 2015

    H Freyja ´Olafsd´ ottir, Caswell Barry, Aman B Saleem, Demis Hassabis, and Hugo J Spiers. Hippocampal place cells construct reward related sequences through unexplored space.Elife, 4:e06063, 2015

    2015

  5. [5]

    The hippocampus as a spatial map: Preliminary evidence from unit activity in the freely-moving rat.Brain Research, 34(1):171–175, 1971

    John O’Keefe and Jonathan Dostrovsky. The hippocampus as a spatial map: Preliminary evidence from unit activity in the freely-moving rat.Brain Research, 34(1):171–175, 1971

    1971

  6. [6]

    Time cells in the hippocampus: a new dimension for mapping memories

    Howard Eichenbaum. Time cells in the hippocampus: a new dimension for mapping memories. Nature Reviews Neuroscience, 15(11):732–744, 2014

    2014

  7. [7]

    Cognitive maps in rats and men.Psychological Review, 55(4):189–208, 1948

    Edward C Tolman. Cognitive maps in rats and men.Psychological Review, 55(4):189–208, 1948

    1948

  8. [8]

    Oxford University Press, 1978

    John O’Keefe and Lynn Nadel.The hippocampus as a cognitive map. Oxford University Press, 1978. 9

    1978

Show all 59 references
  1. [9]

    Simple memory: a theory for archicortex.Philosophical Transactions of the Royal Society of London B, 262(841):23–81, 1971

    David Marr. Simple memory: a theory for archicortex.Philosophical Transactions of the Royal Society of London B, 262(841):23–81, 1971

    1971

  2. [10]

    Computational analysis of the role of the hippocampus in memory.Hippocampus, 4(3):374–391, 1994

    Alessandro Treves and Edmund T Rolls. Computational analysis of the role of the hippocampus in memory.Hippocampus, 4(3):374–391, 1994

    1994

  3. [11]

    James L McClelland, Bruce L McNaughton, and Randall C O’Reilly. Why there are comple- mentary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory.Psychological Review, 102(3):419–457, 1995

    1995

  4. [12]

    The hippocampus as a predictive map.Nature neuroscience, 20(11):1643–1653, 2017

    Kimberly L Stachenfeld, Matthew M Botvinick, and Samuel J Gershman. The hippocampus as a predictive map.Nature neuroscience, 20(11):1643–1653, 2017

    2017

  5. [13]

    The successor representation in human reinforcement learning

    Ida Momennejad, Evan M Russek, Jin H Cheong, Matthew M Botvinick, Nathaniel Douglass Daw, and Samuel J Gershman. The successor representation in human reinforcement learning. Nature human behaviour, 1(9):680–692, 2017

    2017

  6. [14]

    The brain hierarchically rep- resents the past and future during multistep anticipation.Nature Communications, 15(1):9094, 2024

    Hannah Tarder-Stoll, Christopher Baldassano, and Mariam Aly. The brain hierarchically rep- resents the past and future during multistep anticipation.Nature Communications, 15(1):9094, 2024

    2024

  7. [15]

    The hippocampus and inferential reasoning: Building memories to navigate future decisions.Frontiers in Human Neuroscience, 6:70, 2012

    Dagmar Zeithamova, Margaret L Schlichting, and Alison R Preston. The hippocampus and inferential reasoning: Building memories to navigate future decisions.Frontiers in Human Neuroscience, 6:70, 2012

    2012

  8. [16]

    The tolman-eichenbaum machine: unifying space and relational memory through generalization in the hippocampal formation.Cell, 183(5):1249– 1263, 2020

    James CR Whittington, Timothy H Muller, Shirley Mark, Guifen Chen, Caswell Barry, Neil Burgess, and Timothy EJ Behrens. The tolman-eichenbaum machine: unifying space and relational memory through generalization in the hippocampal formation.Cell, 183(5):1249– 1263, 2020

    2020

  9. [17]

    What is a cognitive map? organizing knowledge for flexible behavior.Neuron, 100(2):490–509, 2018

    Timothy EJ Behrens, Timothy H Muller, James CR Whittington, Shirley Mark, Alon B Baram, Kimberly L Stachenfeld, and Zeb Kurth-Nelson. What is a cognitive map? organizing knowledge for flexible behavior.Neuron, 100(2):490–509, 2018

    2018

  10. [18]

    Semantic search as pattern completion across a concept

    Sarah H Solomon and Anna C Schapiro. Semantic search as pattern completion across a concept. Trends in cognitive sciences, 24(2):95–98, 2020

    2020

  11. [19]

    A survey on image data augmentation for deep learning.Journal of Big Data, 6(1):1–48, 2019

    Connor Shorten and Taghi M Khoshgoftaar. A survey on image data augmentation for deep learning.Journal of Big Data, 6(1):1–48, 2019

    2019

  12. [20]

    Generalization through the recurrent interaction of episodic memories: a model of the hippocampal system.Psychological review, 119(3):573, 2012

    Dharshan Kumaran and James L McClelland. Generalization through the recurrent interaction of episodic memories: a model of the hippocampal system.Psychological review, 119(3):573, 2012

    2012

  13. [21]

    A unifying account of replay as context-driven memory reactivation.eLife, 13:RP99931, 2026

    Zhenglong Zhou, Michael J Kahana, and Anna C Schapiro. A unifying account of replay as context-driven memory reactivation.eLife, 13:RP99931, 2026

    2026

  14. [22]

    Adaptive compression as a unifying framework for episodic and semantic memory.Nature Reviews Psychology, pages 1–15, 2025

    David G Nagy, Gerg˝ o Orb´ an, and Charley M Wu. Adaptive compression as a unifying framework for episodic and semantic memory.Nature Reviews Psychology, pages 1–15, 2025

    2025

  15. [23]

    Dinov2: Learning robust visual features without supervision.Transactions on Machine Learning Research, 2024

    Maxime Oquab, Timoth´ ee Darcet, Th´ eo Moutakanni, Huy Vo, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision.Transactions on Machine Learning Research, 2024. 10

    2024

  16. [24]

    VGGT: Visual geometry grounded transformer

    Jianyuan Wang, Minghao Chen, Nikita Karaev, Andrea Vedaldi, Christian Rupprecht, and David Novotny. VGGT: Visual geometry grounded transformer. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 5294–5306, 2025

    2025

  17. [25]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. InInternational conference on machine learning, pag...

    2021

  18. [26]

    Getting aligned on repre- sentational alignment.Transactions on Machine Learning Research, 2025

    Ilia Sucholutsky, Lukas Muttenthaler, Adrian Weller, Andi Peng, Andreea Bobu, Been Kim, Bradley C Love, Christopher J Cueva, Erin Grant, Iris Groen, et al. Getting aligned on repre- sentational alignment.Transactions on Machine Learning Research, 2025

    2025

  19. [27]

    Human alignment of neural network representations

    Lukas Muttenthaler, Jonas Dippel, Lorenz Linhardt, Robert A Vandermeulen, and Simon Ko- rnblith. Human alignment of neural network representations. InThe Eleventh International Conference on Learning Representations, 2023

    2023

  20. [28]

    Human-level 3d shape perception emerges from multi-view learning.arXiv preprint, 2024

    Tyler Bonnen, Jitendra Malik, and Angjoo Kanazawa. Human-level 3d shape perception emerges from multi-view learning.arXiv preprint, 2024. Nature, under review

    2024

  21. [29]

    A is B” fail to learn “B is A

    Lukas Berglund, Meg Tong, Maximilian Kaufmann, Mikita Balesni, Asa Cooper Stickland, Tomasz Korbak, and Owain Evans. The Reversal Curse: LLMs trained on “A is B” fail to learn “B is A”.International Conference on Learning Representations, 2024. *This paper demonstrates the “Re...

    2024

  22. [30]

    Reverse training to nurse the reversal curse

    Olga Golovneva, Zeyuan Allen-Zhu, Jason E Weston, and Sainbayar Sukhbaatar. Reverse training to nurse the reversal curse. InFirst Conference on Language Modeling, 2024

    2024

  23. [31]

    Closing the data-efficiency gap between autoregressive and masked diffusion llms.arXiv preprint arXiv:2510.09885, 2025

    Xu Pan, Ely Hahami, Jingxuan Fan, Ziqian Xie, and Haim Sompolinsky. Closing the data-efficiency gap between autoregressive and masked diffusion llms.arXiv preprint arXiv:2510.09885, 2025

    2025 arXiv

  24. [32]

    On the generalization of language models from in-context learning and finetuning: a controlled study

    Andrew K Lampinen, Arslan Chaudhry, Stephanie CY Chan, Cody Wild, Diane Wan, Alex Ku, J¨ org Bornschein, Razvan Pascanu, Murray Shanahan, and James L McClelland. On the generalization of language models from in-context learning and finetuning: a controlled study. Foundations o...

    2025

  25. [33]

    Deductive closure training of language models for coherence, accuracy, and updatability

    Afra Feyza Aky¨ urek, Ekin Aky¨ urek, Leshem Choshen, Derry Tanti Wijaya, and Jacob Andreas. Deductive closure training of language models for coherence, accuracy, and updatability. In Findings of the Association for Computational Linguistics: ACL 2024, pages 9802–9818, 2024

    2024

  26. [34]

    Syn- thetic continued pretraining

    Zitong Yang, Neil Band, Shuangping Li, Emmanuel Candes, and Tatsunori Hashimoto. Syn- thetic continued pretraining. InThe Thirteenth International Conference on Learning Repre- sentations, 2025. **This paper argues that language models struggle to learn facts about entities an...

    2025

  27. [35]

    New News: System-2 Fine-tuning for Robust Integration of New Knowledge.arXiv preprint arXiv:2505.01812, 2025

    Core Francisco Park, Zechen Zhang, and Hidenori Tanaka. New News: System-2 Fine-tuning for Robust Integration of New Knowledge.arXiv preprint arXiv:2505.01812, 2025

    2025

  28. [36]

    Data-efficient pre-training by scaling synthetic megadocs.arXiv preprint arXiv:2603.18534, 2026

    Konwoo Kim, Suhas Kotha, Yejin Choi, Tatsunori Hashimoto, Nick Haber, and Percy Liang. Data-efficient pre-training by scaling synthetic megadocs.arXiv preprint arXiv:2603.18534, 2026

    2026

  29. [37]

    Princeton University Press, 1954

    Pierre Duhem.The Aim and Structure of Physical Theory. Princeton University Press, 1954. Originally published 1906

    1954

  30. [38]

    Edward Arnold, London, 1960

    Giles S Brindley.Physiology of the Retina and Visual Pathway. Edward Arnold, London, 1960

    1960

  31. [39]

    How computational modeling can force theory building in psychological science.Perspectives on Psychological Science, 16(4):789–802, 2021

    Olivia Guest and Andrea E Martin. How computational modeling can force theory building in psychological science.Perspectives on Psychological Science, 16(4):789–802, 2021

    2021

  32. [40]

    Using goal-driven deep learning models to understand sensory cortex.Nature Neuroscience, 19(3):356–365, 2016

    Daniel LK Yamins and James J DiCarlo. Using goal-driven deep learning models to understand sensory cortex.Nature Neuroscience, 19(3):356–365, 2016

    2016

  33. [41]

    When the ventral visual stream is not enough: A deep learning account of medial temporal lobe involvement in perception

    Tyler Bonnen, Daniel LK Yamins, and Anthony D Wagner. When the ventral visual stream is not enough: A deep learning account of medial temporal lobe involvement in perception. Neuron, 109(17):2755–2766, 2021

    2021

  34. [42]

    Latent learning: episodic memory complements parametric learning by enabling flexible reuse of experiences.Transactions on Machine Learning Research, 2026

    Andrew Kyle Lampinen, Martin Engelcke, Yuxuan Li, Arslan Chaudhry, and James L McClel- land. Latent learning: episodic memory complements parametric learning by enabling flexible reuse of experiences.Transactions on Machine Learning Research, 2026. **This paper argues for an o...

    2026

  35. [43]

    Improving latent generalization using test-time compute.3rd Conference on Language Modeling, 2026

    Arslan Chaudhry, Sridhar Thiagarajan, and Andrew Lampinen. Improving latent generalization using test-time compute.3rd Conference on Language Modeling, 2026. **This paper illustrates how RL can substitute for part of the benefit of retrieval in online augmentation; instead of ...

    2026

  36. [44]

    Chain-of-visual-thought: Teaching vlms to see and think better with continuous visual tokens.arXiv preprint arXiv:2511.19418, 2025

    Yiming Qin, Bomin Wei, Jiaxin Ge, Konstantinos Kallidromitis, Stephanie Fu, Trevor Dar- rell, and Xudong Wang. Chain-of-visual-thought: Teaching vlms to see and think better with continuous visual tokens.arXiv preprint arXiv:2511.19418, 2025

    2025

  37. [45]

    Perception tokens enhance visual reasoning in multimodal language mod- els.arXiv preprint arXiv:2412.03548, 2024

    Mahtab Bigverdi, Zelun Luo, Cheng-Yu Hsieh, Ethan Shen, Dongping Chen, Linda G Shapiro, and Ranjay Krishna. Perception tokens enhance visual reasoning in multimodal language mod- els.arXiv preprint arXiv:2412.03548, 2024

    2024 arXiv

  38. [46]

    Flexible weighting of diverse inputs makes hip- pocampal function malleable.Neuroscience Letters, 680:13–22, 2018

    Mariam Aly and Nicholas B Turk-Browne. Flexible weighting of diverse inputs makes hip- pocampal function malleable.Neuroscience Letters, 680:13–22, 2018

    2018

  39. [47]

    Human replay spontaneously reorganizes experience.Cell, 178(3):640–652, 2019

    Yunzhe Liu, Raymond J Dolan, Zeb Kurth-Nelson, and Timothy EJ Behrens. Human replay spontaneously reorganizes experience.Cell, 178(3):640–652, 2019

    2019

  40. [48]

    Schema representa- tions in distinct brain networks support narrative memory during encoding and retrieval.elife, 11:e70445, 2022

    Rolando Mas ´ ıs-Obando, Kenneth A Norman, and Christopher Baldassano. Schema representa- tions in distinct brain networks support narrative memory during encoding and retrieval.elife, 11:e70445, 2022

    2022

  41. [49]

    Overlapping memory replay during sleep builds cognitive schemata.Trends in cognitive sciences, 15(8):343–351, 2011

    Penelope A Lewis and Simon J Durrant. Overlapping memory replay during sleep builds cognitive schemata.Trends in cognitive sciences, 15(8):343–351, 2011. 12

    2011

  42. [50]

    Generative replay un- derlies compositional inference in the hippocampal-prefrontal circuit.Cell, 186(22):4885–4897, 2023

    Philipp Schwartenbeck, Alon Baram, Yunzhe Liu, Shirley Mark, Timothy Muller, Raymond Dolan, Matthew Botvinick, Zeb Kurth-Nelson, and Timothy Behrens. Generative replay un- derlies compositional inference in the hippocampal-prefrontal circuit.Cell, 186(22):4885–4897, 2023

    2023

  43. [51]

    Human hippocampal ripples coordinate planning sequences and compositional representations in neocortex.Nature Neuroscience, pages 1–11, 2026

    Li He, Xiongfei Wang, Jinbo Zhang, Zhibing Xiao, Xiangyu Hu, Philipp Schwartenbeck, Jacob Bakermans, Tim Behrens, and Yunzhe Liu. Human hippocampal ripples coordinate planning sequences and compositional representations in neocortex.Nature Neuroscience, pages 1–11, 2026

    2026

  44. [52]

    Searching for semantic knowledge: A vector space semantic analysis of the feature generation task.Frontiers in human neuroscience, 13:341, 2019

    Rebecca A Cutler, Melissa C Duff, and Sean M Polyn. Searching for semantic knowledge: A vector space semantic analysis of the feature generation task.Frontiers in human neuroscience, 13:341, 2019

    2019

  45. [53]

    What learning systems do intelligent agents need? complementary learning systems theory updated.Trends in Cognitive Sciences, 20(7):512–534, 2016

    Dharshan Kumaran, Demis Hassabis, and James L McClelland. What learning systems do intelligent agents need? complementary learning systems theory updated.Trends in Cognitive Sciences, 20(7):512–534, 2016

    2016

  46. [54]

    Reactivation of hippocampal ensemble memories during sleep.Science, 265(5172):676–679, 1994

    Matthew A Wilson and Bruce L McNaughton. Reactivation of hippocampal ensemble memories during sleep.Science, 265(5172):676–679, 1994

    1994

  47. [55]

    Reverse replay of behavioural sequences in hippocampal place cells during the awake state.Nature, 440(7084):680–683, 2006

    David J Foster and Matthew A Wilson. Reverse replay of behavioural sequences in hippocampal place cells during the awake state.Nature, 440(7084):680–683, 2006

    2006

  48. [56]

    Odor cues during slow-wave sleep prompt declarative memory consolidation.Science, 315(5817):1426–1429, 2007

    Bj¨ orn Rasch, Christian B¨ uchel, Steffen Gais, and Jan Born. Odor cues during slow-wave sleep prompt declarative memory consolidation.Science, 315(5817):1426–1429, 2007

    2007

  49. [57]

    Strengthening individual memories by reactivating them during sleep.Science, 326(5956):1079–1079, 2009

    John D Rudoy, Joel L Voss, Carmen E Westerberg, and Ken A Paller. Strengthening individual memories by reactivating them during sleep.Science, 326(5956):1079–1079, 2009

    2009

  50. [58]

    Hannah Tarder-Stoll, Christopher Baldassano, and Mariam Aly. The hippocampus rapidly inte- grates sequence representations during novel multistep predictions.Philosophical Transactions of the Royal Society B: Biological Sciences, 381(1954):20250237, 07 2026

    1954

  51. [59]

    Replay comes of age.Annual Review of Neuroscience, 40:581–602, 2017

    David J Foster. Replay comes of age.Annual Review of Neuroscience, 40:581–602, 2017

    2017

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.