Pith. sign in

REVIEW 5 major objections 6 minor 48 references

Narrative Memory in Machines: Multi-Agent Arc Extraction in Serialized TV

T0 review · 5 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A nine-agent system reads episode summaries and stores them as narrative memory, extracting self-contained TV story arcs with 89.3% precision and character entities with 61/62 accuracy.

desk verdict A candid but incremental systems paper: the memory framing is a re-labeling of standard RAG, the evaluation is precision-only with no recall or baseline, yet it's honest and worth a referee. read the letter →

arxiv 2508.07010 v1 pith:65RW54PU submitted 2025-08-09 cs.MM cs.HCcs.MA

Roberto Balestri , Guglielmo Pescatore This is my paper · ORCID
classification cs.MMcs.HCcs.MA
keywords multi-agentsystemsnarrativearcsserializedtelevisionepisodicmemorysemanticworkingLLMvectordatabase
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

This paper is trying to show that a multi-agent system can do something human viewers do across a whole season: turn scattered episode events into remembered story arcs. It builds the system around the human memory analogy—an LLM supplies general narrative knowledge, a vector database stores specific episode events, and a sequence of nine agents acts as working memory that consolidates the two. Tested on the first season of Grey's Anatomy, the system found self-contained 'anthology' arcs with 89.3% precision (25 of 28) and identified 61 of 62 character entities correctly. Its measured weakness is exactly where long-term consolidation should matter: overlapping relationship arcs get split, merged, or missed, and the paper attributes that to the thinness of the episode summaries it is fed. A sympathetic reader would take the paper's claim to be that this memory architecture is a workable, human-refinable scaffold for computational narratology, not a finished replica of human understanding.

What carries the argument

The load-bearing mechanism is a three-part computational memory architecture: an LLM as semantic memory (general narrative conventions and world knowledge), a vector database of arc, progression, and utterance embeddings as episodic memory, and the sequential nine-agent pipeline as working memory that retrieves, compares, merges, verifies, and finally stores narrative information. The atomic unit of episodic memory is the utterance—a single event or interaction within a progression—so retrieval and reasoning happen at event granularity, with each progression grouping an episode's events under one arc and each arc typed as Anthology, Soap, or Genre-Specific.

What would settle it

Run the identical nine-agent pipeline on another serialized drama's first season using only official short episode synopses instead of fan-written summaries, and measure anthology-arc precision: if it falls well below the reported 89.3%, the capability is an artifact of summary richness rather than the memory architecture. Alternatively, feed the same Grey's Anatomy season subtitles plus scene descriptions and check whether the missed Meredith–Derek relational arc is recovered.

Watch

Extended reading notes

Core claim

The central discovery is a working demonstration that narrative arc extraction can be organized as memory management rather than pure classification. The system treats each cleaned episode summary as input to a nine-agent pipeline; agents identify existing arcs, extract new anthology, soap, and genre-specific arcs, optimize and deduplicate them, enhance details, verify progressions and character roles, and finally commit verified arcs to a vector database as episodic traces made of utterance-level memory units. The evaluation against a human scholar's gold standard found the system strong where memory demands are light—self-contained anthology arcs and named character entities—and weak where

Load-bearing premise

The whole evaluation rests on the assumption that fan-written episode summaries contain enough explicit narrative detail for multi-episode arcs to be recognized; the authors concede that when summaries do not make a relationship salient, no memory mechanism in the system can recover it.

Editorial extensions

If this is right

  • If the architecture works as claimed, serialized written formats (novels, episodic web fiction, comics) should be analyzable by the same pipeline with no multimodal gap, since all narrative information is in the text.
  • The measured split between strong anthology-arc precision and weaker relational-arc consolidation implies near-term automation is safest for discrete, self-contained storylines and entity extraction, while long-range relationship arcs still need human review.
  • Because verified arcs are committed to a vector database, the output is a durable, queryable narrative memory: users can filter by arc type, character, or episode and explore semantic clusters.
  • The human-editing interface turns the system's memory into a revisable object: analysts can merge duplicate arcs, split over-broad ones, and regenerate progressions, which should cumulatively improve downstream retrieval.
  • Adding subtitles or scene descriptions to the episodic store is a direct next step implied by the failure analysis: richer inputs should supply the relational salience the current summaries omit.

Reading between the lines

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

  • The precision numbers come from one season of one series; if the architecture is as general as the memory framing suggests, similar precision on self-contained arcs should reproduce on other serialized dramas with comparable paratext quality.
  • The failure pattern predicts a testable ordering: coarser summaries (official one-line synopses) should worsen relational-arc consolidation more than anthology-arc precision, since anthology arcs depend less on cross-episode memory.
  • A cheap fix suggested by the results, not proposed by the paper, is a second verification pass over characters and arcs analogous to the existing Final Reviewer agent, which would likely remove duplicate entities like 'Frost' and 'Jerry Frost' at some compute cost.
  • The paper's own framing implies that narrative forgetting could be modeled deliberately—weighting older progressions lower or decaying embeddings—turning the system's memory from a perfect store into a salience-shaped one closer to human recall.
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

5 major / 6 minor

Summary. The paper presents a multi-agent system (MAS) for extracting narrative arcs from serialized TV episode summaries, conceptualized through analogies to human declarative memory (episodic, semantic, working). The architecture uses an LLM (GPT-4o) as semantic memory, a Chroma vector database as episodic memory, and a nine-agent sequential workflow as working memory. The system is tested on Season 1 of Grey's Anatomy using fan-wiki episode summaries, with a graphical interface for human refinement. Quantitative results are reported only for Anthology arcs (precision 89.3%, 25/28) and character entities (61/62 correct). Soap and Genre-Specific arcs, which require cross-episode memory consolidation, are evaluated qualitatively, with acknowledged failures such as duplicated Izzie Stevens arcs, a missed Meredith/Derek relational arc, and a misclassified 'Roommates Dynamics' arc. The authors candidly state that the system's performance depends on the detail level of the input paratexts and that the memory architecture is a functional analogue, not a replication of human memory.

Significance. If the reported precision values were supported by recall, F1, and baseline comparisons, the paper would offer a useful proof-of-concept for memory-inspired multi-agent narrative analysis, with open-source code and a human-in-the-loop refinement tool. However, the current evaluation limits the strength of the claims. The system is a plausible prototype, but the central claim that it can extract and consolidate narrative arcs in serialized TV remains insufficiently validated. The paper's honest reporting of limitations and explicit memory-based framing are strengths, as is the availability of the repository.

major comments (5)
  1. [Section 9.2] The quantitative anchor is precision-only for Anthology arcs (25/28) and character entities (61/62). No recall, F1, or total number of gold-standard arcs is reported. High precision without recall could correspond to extracting only a small, easy subset of the true arcs, so the central claim that the system is a 'workable pipeline' for structuring serialized narrative is not established. Recall and F1 must be computed by comparing against a full gold-standard arc set.
  2. [Section 9.2] Soap and Genre-Specific arcs are the types that actually exercise the multi-episode memory consolidation that the paper emphasizes, yet they receive no quantitative evaluation. The three named failures (duplicated Izzie Stevens arcs, missed Meredith/Derek relational arc, misclassified 'Roommates Dynamics') are illustrative but provide no basis for assessing precision/recall for these types. Per-type metrics must be reported, especially for Soap arcs, to support the memory-consolidation claims.
  3. [Section 9.1] The gold standard is a single human scholar's analysis, with no inter-annotator agreement and no explicit protocol for matching system arcs to human arcs. Arc boundaries and taxonomy are subjective; a single-rater gold standard makes the reported precision difficult to interpret and reproduce. Provide a detailed matching criterion, ideally with multiple annotators, or at least a full listing of gold-standard arcs.
  4. [Section 9 (overall)] No baseline comparison is provided. The paper claims the memory-based multi-agent architecture is beneficial, but there is no comparison to a zero-shot LLM prompted to extract arcs from the same summaries, or to the prior system in [3]. Without a baseline, the added complexity of the nine-agent architecture and the vector database is not justified. At least one simple baseline should be evaluated in the same setting.
  5. [Sections 7.3 and 8.2] The deduplication and merging of arcs rely on semantic-similarity thresholds and a Jaccard similarity threshold for character merging, but no sensitivity analysis is provided. Different thresholds could materially change the outcomes, particularly for Soap/Genre arcs, where failures concentrate. Report the exact threshold values used and test robustness to them.
minor comments (6)
  1. [Section 8.2] Typo: 'mnarrative arcs' should be 'narrative arcs'.
  2. [Section 8.2] The text 'ArcCreationandEditing' is missing spaces; should be 'Arc Creation and Editing'.
  3. [Section 6.1] The license statement for the fan wiki is brief; consider adding details about how the content was reused and attribution.
  4. [Section 5] The GPT-4o model version and access date are not specified; adding a model snapshot would improve reproducibility.
  5. [Section 9.2] The phrase '62 character entities' is ambiguous—clarify whether this is unique characters or character mentions across the season.
  6. [Section 7] The exact prompt templates for each agent are not included; an appendix with prompts would strengthen reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the arc-extraction claims are evaluated against an external human gold standard, and no fitted parameter or self-citation chain makes the results equivalent to the inputs.

full rationale

The paper's central claim is that the multi-agent system extracts narrative arcs from episode summaries. The reported quantitative results (89.3% precision on Anthology arcs, 61/62 correct character entities) are computed by comparing the system's outputs against a human scholar's analysis of the actual episodes, which is an external reference standard. No parameter is fitted to that gold standard and then re-reported as a prediction; the numbers are direct counts. The architecture builds on the authors' prior work [3] and the arc taxonomy draws on [28,34], but these self-citations are antecedent design choices, not the evidence for the load-bearing claim. The taxonomy is imposed on the LLM as a classification scheme, not derived from the taxonomy as a prediction. The paper explicitly flags that the system's memory depends on 'the detail level of the input paratexts' (Sec. 9.2), which is an input-channel limitation rather than a circular step. Missing recall/F1 and baselines are correctness-validity concerns, not identity-by-construction. No equation or definition in the paper reduces a prediction to an input, so no circular step can be identified.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The paper fits no numerical constants to data. Its commitments are taxonomic (three arc types imposed on the LLM), procedural (nine agents, LLM-based cleaning, embedding pipeline), and architectural (off-the-shelf components relabeled as memory systems). The load-bearing assumptions are the input-channel sufficiency of paratexts, the validity of an undisclosed single human rater as gold standard, and the adopted taxonomy. Free parameters are limited to unreported similarity thresholds that gate deduplication and entity merging. No new physical or formal entities are postulated; the 'utterance as atomic memory trace' is an analytical granularity choice without independent falsifiable handles.

free parameters (2)
  • Semantic similarity decision threshold for arc linking and deduplication
    Sections 7.1 and 7.3 retrieve similar arcs and let the LLM decide if they are the same storyline; no numeric cutoff is reported, so the effective threshold is an unspecified engineering choice embedded in prompts.
  • Jaccard similarity threshold for duplicate character suggestions
    Section 8.2 uses a Jaccard-index threshold to flag duplicate characters in the interface; the value is unreported and peripheral to the central arc-extraction claim.
assumptions (4)
  • domain assumption Fan-wiki episode summaries, after LLM simplification, contain sufficient narrative information to reconstruct multi-episode arcs.
    The whole episodic store is built from paratexts (Sections 6.1, 6.2, 9.1). The paper's own failures, the missed Meredith/Derek arc and the over-merged Roommates arc, are attributed to this channel (Section 9.2), so the premise is load-bearing and only partially satisfied.
  • domain assumption A single human analyst's arc annotations constitute a valid gold standard.
    Section 9.1 treats watching each episode at least twice as ground truth; the rater's identity, independence from the authors, and inter-rater agreement are undisclosed, so the evaluation's validity rests on this unstated premise.
  • ad hoc to paper The three-way arc taxonomy (Anthology, Soap, Genre-Specific) is the right organizing scheme for serial narrative.
    Section 4 derives the types from [28,34], prior papers by the same research group, and the LLM agents are instructed to classify into exactly these types, so the discovered structure is partly imposed by the taxonomy.
  • domain assumption The cognitive memory mapping (LLM as semantic memory, vector DB as episodic memory, agent pipeline as working memory) is a faithful model, not just a metaphor.
    Section 3 builds the design on Tulving and Baddeley, but Section 10.2 concedes the system is only a 'functional analogue', so the strength of the theoretical contribution hinges on how literally this mapping is read.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Narrative Memory in Machines: Multi-Agent Arc Extraction in Serialized TV." pith.science (2026). https://pith.science/paper/65RW54PU

@misc{pith2026250807010,
  author       = {Pith},
  title        = {Pith review of: Narrative Memory in Machines: Multi-Agent Arc Extraction in Serialized TV},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/65RW54PU}},
  note         = {Machine review of arXiv:2508.07010}
}
read the original abstract

Serialized television narratives present significant analytical challenges due to their complex, temporally distributed storylines that necessitate sophisticated information management. This paper introduces a multi-agent system (MAS) designed to extract and analyze narrative arcs by implementing principles of computational memory architectures. The system conceptualizes narrative understanding through analogues of human memory: Large Language Models (LLMs) provide a form of semantic memory for general narrative patterns, while a vector database stores specific arc progressions as episodic memories. A multi-agent workflow simulates working memory processes to integrate these information types. Tested on the first season of Grey's Anatomy (ABC 2005-), the MAS identifies three arc types: Anthology (self-contained), Soap (relationship-focused), and Genre-Specific. These arcs and their episodic developments are stored in a vector database, facilitating structured analysis and semantic comparison. To bridge automation with critical interpretation, a graphical interface enables human oversight and refinement of the system's narrative memory. While demonstrating strong performance in identifying Anthology Arcs and character entities, the system's reliance on textual paratexts (episode summaries) revealed limitations in discerning overlapping arcs and opaque dynamics, underscoring the challenges in computational memory consolidation versus human holistic understanding. This memory-centric approach highlights the potential of combining AI-driven memory processing with human expertise. Beyond television, it offers promise for serialized written formats where narrative is entirely text-based. Future work will focus on integrating multimodal inputs to enrich episodic memory, refining memory integration mechanisms within the MAS, and expanding testing across diverse genres.

Figures

Figures reproduced from arXiv: 2508.07010 by the authors.

Figure 1
Figure 1. Diagram illustrating the hierarchical structure of Arcs, Progressions, and Ut￾terances. Summarizing, in our system, the object Narrative Arc includes several fields. Each arc is assigned a unique identifier (arc_id), along with a title and description encapsulating its central theme or conflict. The arcs also contain progressions, which are a list of Progression objects representing [PITH_FULL_IMAGE:figures/full_fi… view at source ↗
Figure 2
Figure 2. Narrative Arc Extraction Process as seen in [3] 7.1 Memory-Informed System Architecture The MAS architecture is inherently memory-informed. Each agent’s op￾eration relies on accessing and manipulating different facets of the sys￾tem’s memory [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. The main view of the graphical interface, displaying narrative arcs and their episodic progressions. As seen in [3] [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: 3D PCA visualizer for clustering arcs based on semantic similarity, offering a view into the episodic memory’s structure. As seen in [3] [PITH_FULL_IMAGE:figures/full_fig_p017_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 43 canonical work pages

  1. [3]

    Multi-Agent System for AI-Assisted Extraction of Narrative Arcs in TV Series

    Balestri, R., Pescatore, G.: Multi-agent system for ai-assisted extrac- tion of narrative arcs in tv series. arXiv preprint arXiv:2503.04817 (2025)

    work page Pith review arXiv 2025

  2. [1]

    In: 2023 15th Inter- national Congress on Advanced Applied Informatics Winter (IIAI- AAI-Winter)

    Aoki, N., Mori, N., OKada, M.: Analysis of llm-based narrative generation using the agent-based simulation. In: 2023 15th Inter- national Congress on Advanced Applied Informatics Winter (IIAI- AAI-Winter). pp. 284–289. IEEE (2023)

    work page 2023

  3. [2]

    Journal of communication disorders36(3), 189–208 (2003)

    Baddeley, A.: Working memory and language: An overview. Journal of communication disorders36(3), 189–208 (2003)

    work page 2003

  4. [4]

    (eds.): Russian Formalism: A Collection of Ar- ticles and Texts in Translation

    Bann, S., Bowlt, J.E. (eds.): Russian Formalism: A Collection of Ar- ticles and Texts in Translation. 20th Century Studies, Scottish Aca- demic Press, Edinburgh (1973), distributed by Chatto and Windus, London

    work page 1973

  5. [5]

    In: Reading Contem- porary Serial Television Universes, pp

    Beveridge, A., Chemers, M.: The game of game of thrones: Net- worked concordances and fractal dramaturgy. In: Reading Contem- porary Serial Television Universes, pp. 201–225. Routledge (2018)

    work page 2018

  6. [6]

    In: 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM)

    Bost, X., Labatut, V., Gueye, S., Linarès, G.: Narrative smoothing: dynamic conversational network for the analysis of tv series plots. In: 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM). pp. 1111–1118. IEEE (2016)

    work page 2016

  7. [7]

    looking for lexical signatures in gomorrah

    Dalla Torre, P., Fantozzi, P., Naldi, M., Pellegriti, P.: 3. looking for lexical signatures in gomorrah. AUDIOVISUAL DATA p. 41 (2023)

    work page 2023

  8. [8]

    In: Avezzù, G., Rocchi, M

    Degli Esposti, M., Pescatore, G.: Exploring tv seriality and television studies through data-driven approaches. In: Avezzù, G., Rocchi, M. (eds.) Audiovisual Data: Data-Driven Perspectives for Media Stud- ies. 13th Media Mutations International Conference, pp. 23–40. Me- dia Mutations Publishing, Bologna (2023). https://doi.org/10.214 28/93b7ef64.ec022085...

    work page 2023

Show all 48 references
  1. [9]

    Literature, Culture, Theory, Cambridge University Press (1997)

    Genette, G.: Paratexts: Thresholds of Interpretation. Literature, Culture, Theory, Cambridge University Press (1997)

    1997

  2. [10]

    arXiv preprint arXiv:2409.05556 (2024)

    Ghafarollahi, A., Buehler, M.J.: Sciagents: Automating scientific discovery through multi-agent intelligent graph reasoning. arXiv preprint arXiv:2409.05556 (2024)

    2024 arXiv

  3. [11]

    Psychological review101(3), 371 (1994)

    Graesser, A.C., Singer, M., Trabasso, T.: Constructing inferences during narrative text comprehension. Psychological review101(3), 371 (1994)

    1994

  4. [12]

    NYU Press (2017)

    Gray, J., Sandvoss, C., Harrington, C.L.: Fandom: Identities and communities in a mediated world. NYU Press (2017)

    2017

  5. [13]

    Haywood, S., Warren, G., Wolf, A.: Understand tokens (May 2024), https://learn.microsoft.com/en-us/dotnet/ai/conceptual/underst anding-tokens, microsoft Learn

    2024

  6. [14]

    IEEE Intelligent Systems and their applica- tions 13(4), 18–28 (1998)

    Hearst, M.A., Dumais, S.T., Osuna, E., Platt, J., Scholkopf, B.: Sup- port vector machines. IEEE Intelligent Systems and their applica- tions 13(4), 18–28 (1998)

    1998

  7. [15]

    MIT press (2017) Multi-Agent Arc Extraction in Serialized TV 23

    Herman, D.: Storytelling and the Sciences of Mind. MIT press (2017) Multi-Agent Arc Extraction in Serialized TV 23

    2017

  8. [16]

    a multidisci- plinary approach to media worlds

    Innocenti, V., Pescatore, G.: Narrative ecosystems. a multidisci- plinary approach to media worlds. In: World Building Transmedia, Fans, Industries, pp. 164–183. Amsterdam University Press (2017)

    2017

  9. [17]

    In: Reading contemporary serial television universes, pp

    Innocenti, V., Pescatore, G.: The evolution of characters in tv series: morphology, selection, and remarkable cases in narrative ecosystems. In: Reading contemporary serial television universes, pp. 93–110. Routledge (2018)

    2018

  10. [18]

    arXiv preprint arXiv:2110.09856 (2021)

    Janosov, M.: Network science predicts who dies next in game of thrones. arXiv preprint arXiv:2110.09856 (2021)

    2021 arXiv

  11. [19]

    Routledge (2012)

    Jenkins, H.: Textual poachers: Television fans and participatory cul- ture. Routledge (2012)

    2012

  12. [20]

    Advances in Neural Information Processing Systems33, 9459–9474 (2020)

    Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W.t., Rocktäschel, T., et al.: Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in Neural Information Processing Systems33, 9459–9474 (2020)

    2020

  13. [21]

    arXiv preprint arXiv:1707.05213 (2017)

    Liu, D., Albergante, L.: Balance of thrones: a network study on game of thrones. arXiv preprint arXiv:1707.05213 (2017)

    2017 arXiv

  14. [22]

    PloS one14(12), e0226025 (2019)

    Min, S., Park, J.: Modeling narrative structure and dynamics with networks, sentiment analysis, and topic modeling. PloS one14(12), e0226025 (2019)

    2019

  15. [23]

    Simon and Schuster (1986)

    Minsky, M.: Society of mind. Simon and Schuster (1986)

    1986

  16. [24]

    Intermediality and storytelling24, 78–98 (2010)

    Mittell, J.: Previously on: Prime time serials and the mechanics of memory. Intermediality and storytelling24, 78–98 (2010)

    2010

  17. [25]

    New York University Press (2015)

    Mittell, J.: Complex TV: The Poetics of Contemporary Television Storytelling. New York University Press (2015)

    2015

  18. [26]

    In: Proceedings of the international multiconference of engineers and computer scien- tists

    Niwattanakul, S., Singthongchai, J., Naenudorn, E., Wanapu, S.: Us- ing of jaccard coefficient for keywords similarity. In: Proceedings of the international multiconference of engineers and computer scien- tists. vol. 1, pp. 380–384 (2013)

    2013

  19. [27]

    In: Cognition, Emotion, and Aesthetics in Contemporary Serial Televi- sion, pp

    Pérez, H.J., Ortiz, M.J.: Multi-plot structure in television serials. In: Cognition, Emotion, and Aesthetics in Contemporary Serial Televi- sion, pp. 47–67. Routledge (2021)

    2021

  20. [28]

    interpre- tative hypotheses and research perspectives

    Pescatore, G., Rocchi, M.: Narration in medical dramas i. interpre- tative hypotheses and research perspectives. La valle dell’eden1, 107–115 (2019)

    2019

  21. [29]

    University of Chicago Press (2019)

    Piper, A.: Enumerations: data and literary study. University of Chicago Press (2019)

    2019

  22. [30]

    Poetics98, 101793 (2023)

    Piper, A., Toubia, O.: A quantitative study of non-linearity in sto- rytelling. Poetics98, 101793 (2023)

    2023

  23. [31]

    Wayne State University Press (2012)

    Propp, V.Y.: The Russian Folktale by Vladimir Yakovlevich Propp. Wayne State University Press (2012)

    2012

  24. [32]

    Arxiv Preprint (2018)

    Radford, A., Narasimhan, K., Salimans, T., Sutskever, I., et al.: Im- proving language understanding by generative pre-training. Arxiv Preprint (2018)

    2018

  25. [33]

    EPJ data science5(1), 1–12 (2016)

    Reagan, A.J., Mitchell, L., Kiley, D., Danforth, C.M., Dodds, P.S.: The emotional arcs of stories are dominated by six basic shapes. EPJ data science5(1), 1–12 (2016)

    2016

  26. [34]

    Humanities and Social Sciences Com- munications 9(1), 1–11 (2022) 24 R

    Rocchi, M., Pescatore, G.: Modeling narrative features in tv series: coding and clustering analysis. Humanities and Social Sciences Com- munications 9(1), 1–11 (2022) 24 R. Balestri and G. Pescatore

    2022

  27. [35]

    Schacter, D.L.: The seven sins of memory: insights from psychology andcognitiveneuroscience.Americanpsychologist 54(3), 182(1999)

    1999

  28. [36]

    Schank, R.C.: A theory of reminding and learning in computers and people (1982)

    1982

  29. [37]

    In: 2015 IEEE International Conference on Big Data (Big Data)

    Schmidt, B.M.: Plot arceology: A vector-space model of narrative structure. In: 2015 IEEE International Conference on Big Data (Big Data). pp. 1667–1672. IEEE (2015)

    2015

  30. [38]

    Proceedings of the National Academy of Sciences93(24), 13515–13522 (1996)

    Squire, L.R., Zola, S.M.: Structure and function of declarative and nondeclarative memory systems. Proceedings of the National Academy of Sciences93(24), 13515–13522 (1996)

    1996

  31. [39]

    Todorov,T.,Weinstein,A.A.:Structuralanalysisofnarrative.Novel: A Forum on Fiction3, 70–76 (1969)

    1969

  32. [40]

    Hippocampus8(3), 198–204 (1998)

    Tulving, E., Markowitsch, H.J.: Episodic and declarative memory: role of the hippocampus. Hippocampus8(3), 198–204 (1998)

    1998

  33. [41]

    Advances in Neural Informa- tion Processing Systems (2017)

    Vaswani, A.: Attention is all you need. Advances in Neural Informa- tion Processing Systems (2017)

    2017

  34. [42]

    Int J Eng Res9 (2020)

    Vimal, B.: Application of logistic regression in natural language pro- cessing. Int J Eng Res9 (2020)

    2020

  35. [43]

    arXiv preprint arXiv:2502.00592 (2025)

    Wang, Y., Krotov, D., Hu, Y., Gao, Y., Zhou, W., McAuley, J., Gutfreund, D., Feris, R., He, Z.: M+: Extending memoryllm with scalable long-term memory. arXiv preprint arXiv:2502.00592 (2025)

    2025 arXiv

  36. [44]

    In: Proceedings of the 58th Annual Meeting of the Association for Computational Lin- guistics

    Wilmot, D., Keller, F.: Modelling suspense in short stories as un- certainty reduction over neural representation. In: Proceedings of the 58th Annual Meeting of the Association for Computational Lin- guistics. pp. 1763–1788. Association for Computational Linguistics, Online (J...

    2020 doi

  37. [45]

    arXiv preprint arXiv:1702.01923 (2017)

    Yin, W., Kann, K., Yu, M., Schütze, H.: Comparative study of cnn and rnn for natural language processing. arXiv preprint arXiv:1702.01923 (2017)

    2017 arXiv

  38. [46]

    ieee Computational intelligenCe magazine13(3), 55–75 (2018)

    Young, T., Hazarika, D., Poria, S., Cambria, E.: Recent trends in deep learning based natural language processing. ieee Computational intelligenCe magazine13(3), 55–75 (2018)

    2018

  39. [47]

    arXiv preprint arXiv:2410.02507 (2024)

    Yuan, W., Cao, J., Jiang, Z., Kang, Y., Lin, J., Song, K., Yan, P., Sun, C., Liu, X., et al.: Can large language models grasp legal theo- ries? enhance legal reasoning with insights from multi-agent collab- oration. arXiv preprint arXiv:2410.02507 (2024)

    2024 arXiv

  40. [48]

    Psychological bul- letin 133(2), 273 (2007)

    Zacks, J.M., Speer, N.K., Swallow, K.M., Braver, T.S., Reynolds, J.R.: Event perception: a mind-brain perspective. Psychological bul- letin 133(2), 273 (2007)

    2007

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.