Pith. sign in

REVIEW 3 major objections 4 minor 44 references

Vision-language models describe the same image with less than ten percent lexical overlap when the observer's goal changes, and the paper argues this context-dependence reflects genuine computation rather than sampling noise.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 23:28 UTC pith:U23T2B6C

load-bearing objection The 90% context-dependence headline is confounded by prompt surface differences and should not be cited as-is, but the paper is transparent, ships code/data, and has a real qualitative finding; it deserves a serious referee. the 3 major comments →

arxiv 2603.04419 v2 pith:U23T2B6C submitted 2026-02-14 cs.CL cs.AIcs.LG

Context-Dependent Affordance Computation in Vision-Language Models

Murad Farzulla This is my paper
classification cs.CL cs.AIcs.LG
keywords vision-language modelsaffordancescontext-dependent processingscene understandingfunctional semanticsJaccard similarityTucker decompositionjust-in-time ontology
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper tries to establish that vision-language models do not produce a stable, context-free description of a visual scene: the same image receives a different functional reading depending on who the model is told is looking. Affordances — the action possibilities a scene offers an agent — are computed in a strongly context-dependent way: across 479 images and seven agent personas, word-level Jaccard similarity (a standard overlap measure between token sets) averaged 0.095 between contexts, meaning more than 90 percent of the lexical scene description shifts with context; sentence-level cosine similarity was 0.415, so about 58.5 percent of semantic content shifts as well. The drift survives replication on a second model and survives stochastic sampling controls, and tensor decomposition reveals stable latent factors such as a culinary-oriented dimension and an access dimension. The authors present this as behavioral evidence for context-dependent affordance computation and draw a design consequence for robotics: query-dependent, just-in-time ontologies rather than static world models, while explicitly declining to claim any processing order.

Core claim

The central discovery is 'affordance drift': the set of objects and uses a vision-language model reports for one image changes massively when the prompt changes the agent's identity or goal, from neutral observer to chef, security professional, child, wheelchair user, urgent survivor, or leisured explorer. The mean word-level Jaccard similarity between context conditions is 0.095 (95% CI [0.093, 0.096], p<0.0001), object-level 0.119, and a replication on a second, architecturally distinct model gives 0.181. Because sentence-embedding cosine similarity is much higher (0.415), the authors conclude that surface vocabulary shifts more than underlying meaning, leaving a semantic invariant core. T

What carries the argument

The carrying mechanism is a controlled comparison: the same image is prompted with seven agentic persona primes; outputs are parsed into object/affordance lists; and pairwise Jaccard similarity over word tokens and object names measures how much of the functional scene ontology changes with context. Stochastic baselines across four temperatures and five seeds separate genuine context effects from sampling noise, and Tucker decomposition — a three-mode factor analysis separating image, context, and embedding dimensions — reveals the latent structure of the shift. The Jaccard coefficient is the load-bearing statistic: the claim that more than 90 percent of lexical scene description is context-

Load-bearing premise

The load-bearing premise is that the seven differently worded prompts really create seven different agent contexts; because the paper has not yet run the same-prompt/different-persona and paraphrase controls it lists as future work, the 90 percent drift could in principle be instruction-following surface variation.

What would settle it

A decisive check is the same-prompt/different-persona control: keep the wording of the prime fixed and vary only the named persona, or vary the wording while keeping the persona fixed. If pairwise Jaccard similarity follows the wording rather than the persona, the 90 percent drift is prompt sensitivity, not agent context; if it follows the persona, the paper's interpretation survives.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • If the drift is real, vision pipelines that first build a context-free geometric representation are spending most of their computation on the roughly 10 percent of functional signal that is invariant across tasks.
  • The lexical-versus-semantic gap (90.5% vs 58.5% context-dependent) implies that scene meaning retains a substantial invariant core even as vocabulary changes, so context re-weights salience rather than creating or destroying affordances.
  • The stable orthogonal factors (a culinary manifold and an access axis) mean context-dependent affordance structure is organized around functional categories, not diffuse variation.
  • The effect replicates across two architecturally distinct vision-language models, suggesting it is a general property of models trained on naturalistic image-text data.
  • For embodied AI, the result motivates just-in-time ontology: construct task-specific functional projections at query time rather than maintaining a single static world model, though the paper labels this a suggested direction rather than a proven design.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Extension: the paper's own list of remaining controls — same prompt with different personas, and prompt paraphrase — is the decisive test; if those controls show low drift, the 90% figure is surface prompt sensitivity, not agentic context.
  • Extension: the roughly 40% semantic overlap that survives context shifts suggests a hybrid architecture: a small stable geometric core plus a large dynamic task projection, rather than the pure just-in-time alternative the paper sketches.
  • Extension: the access axis separating child from wheelchair-user contexts offers a concrete testbed for the paper's untested action-distance prediction: compare spatial reasoning or navigation judgments under the two personas and see whether behavior tracks Euclidean or action-based distance.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This manuscript characterizes context-dependent affordance computation in vision-language models (VLMs). Using 479 COCO images and seven persona prompts (neutral, chef, security, child, wheelchair-user, urgent, leisure), it asks Qwen3-VL-30B-A3B to output JSON affordance descriptions and computes pairwise Jaccard similarity across primes. The mean word-level Jaccard of 0.095 is interpreted as '>90% of lexical scene description is context-dependent'; sentence-level cosine similarity of 0.415 is interpreted as 58.5% context-dependence. A LLaVA-1.5-13B replication, stochastic sampling controls, and Tucker decomposition of sentence embeddings are reported, and the paper proposes a 'semantic-first' processing architecture and Just-In-Time Ontology for robotics. The paper explicitly limits its claims to behavioral output and lists several untested controls.

Significance. The qualitative phenomenon—that a VLM's reported objects and affordances shift when the user prompt changes—is plausible and consistent with prior work on instruction-following and task-conditioned affordance ranking. The paper has notable strengths: public code/data, a large pairwise sample, cross-model replication, stochastic baselines, bootstrap stability analysis, and unusually explicit statements about what is not claimed. If the 90% magnitude were valid, it would be an important quantitative anchor for debates about static versus dynamic scene representations in robotics and cognitive AI. However, the central magnitude is not currently established: the primes differ in surface task demands, the key same-prompt/paraphrase controls are acknowledged as future work, and 1−Jaccard is not a variance-explained quantity. The headline contribution therefore requires substantial additional support.

major comments (3)
  1. [§4.4, §6.6, Appendix A.1 (Table 12)] The headline inference from J̄=0.095 to '>90% of functional scene ontology is context-dependent' requires that the seven primes isolate agent context. They do not. P0 asks for the 3 most prominent objects, geometric properties, and standard functions; P1–P4 ask for 3 role-specific critical/interesting objects; P5 asks for a single survival tool; P6 asks what 'catches your eye.' Word-level Jaccard (Eq. 8) is computed over token sets of concatenated name+affordance+reasoning, so changes in requested object count, output length, and instruction vocabulary directly deflate J; P5's single-object output alone caps overlap with three-object outputs near 1/3. The paper itself lists same-prompt/different-image and prompt-paraphrase sensitivity as 'Remaining Controls for Future Work' (§4.4, §6.6), i.e., the control that would separate agent context from prompt wording is absent. As it stands, the
  2. [§5.6, Table 10] The paper treats 1−J and 1−cosine as 'proportion of variance explained' and as 'Ctx-Dep.' percentages. Jaccard is a set-overlap coefficient; its complement is not a variance decomposition and cannot be interpreted as 'context explains approximately 90% of the variance.' The same holds for cosine similarity: 0.415 similarity does not imply 58.5% of variance is context-dependent. The only variance-explained quantity in the paper is the Tucker decomposition's 46.6%, and that is tensor-approximation variance, not a decomposition of the Jaccard contrast. This is a category error in the central statistic.
  3. [§5.4, §6.4, Table 7] Table 7 explicitly defines 'Var. %' as each factor's share of captured variance (summing to 100% of explained variance). Yet §5.4 says Dim1 'accounts for only 0.9% of variance,' and §6.4 cites 'Tucker Dim1's low variance contribution (0.9%)' to argue geometry-invariant features are a small residual. Because only 46.6% of total tensor variance is explained, Dim1's 0.9% is a share of that captured variance, not of total variance. The architectural inference drawn from this number is therefore not supported by the reported decomposition.
minor comments (4)
  1. [Abstract vs. Table 3 / §5.7] The abstract reports LLaVA mean J=0.160, whereas Table 3 reports 0.1807; the abstract reports 2,384 stochastic inference runs, while §5.7 describes 7,000 runs. Please harmonize these numbers.
  2. [§6.6 vs. §5.8] §6.6 refers to 'Lemmatized Jaccard' as an alternative metric, but §5.8/Table 10 implements stopword-filtered Jaccard, not lemmatization; Section 4.3.4 explicitly states that no lemmatization/stemming is used.
  3. [§5.3] The Visual Genome human baseline is not collected under goal priming, so it cannot validate context-dependent extraction; the comparison is suggestive only and should be labeled as such.
  4. [§4.4, Table 2] The null threshold J<0.5 is arbitrary; since the observed J=0.095 is within 2–9× of the paper's own random-baseline estimate (J≈0.01–0.05, §4.4), a more informative benchmark would be random-pair or same-prompt overlap.

Circularity Check

0 steps flagged

No significant circularity: the reported drift is an output-contrast measurement with acknowledged construct-validity limits, not a derivation from its own inputs.

full rationale

The paper's headline '>90% context-dependent' is the complement of the observed pairwise Jaccard similarity between model outputs under seven primes (Eq. 8). This is an operational measurement, not a fitted parameter renamed as a prediction. No equation reduces a target result to an input by construction. H2 is tested by comparing outputs, and the paper explicitly limits the claim to behavioral output ('We do not claim to establish processing order or architectural primacy', Section 7). The only self-citation (Farzulla, 2025) appears as a non-load-bearing aside linking theta_history to developmental training, not as support for the drift result. The acknowledged missing controls in Section 4.4 ('Same-prompt, different-image' and 'Prompt paraphrase sensitivity') and the metric-dependence caveat in Section 5.8 (cosine similarity yields 58.5% context-dependence) are construct-validity and confound concerns, not circular reductions. The Tucker 'Culinary Manifold' and 'Access Axis' labels are post-hoc interpretations of stable factor loadings, not predictions derived from those loadings. Thus no specific circular step can be exhibited, and the score is 0.

Axiom & Free-Parameter Ledger

2 free parameters · 4 axioms · 0 invented entities

No fitted parameters are used in a derivation, but the primary temperature and Tucker rank are hand-chosen and affect the headline numbers. The main interpretive load is carried by two ad-hoc statistical assumptions: the 0.5 null threshold and the equation of 1-J with context-dependent proportion. The 'Culinary Manifold' and 'Access Axis' are data-derived factor labels, not invented entities.

free parameters (2)
  • Primary inference temperature = 0.7 (primary; 0.0-1.0 in stochastic baseline)
    Chosen by hand as a 'principled trade-off' (Section 4.1); affects output variability and therefore all measured overlap values.
  • Tucker rank = [10,3,10]
    Selected after comparing ranks [5,3,5], [10,3,10], [15,3,15], [20,3,20] (Section 5.9); determines interpretable factors and explained variance.
axioms (4)
  • ad hoc to paper The threshold J<0.5 defines 'more difference than overlap' and thus context-dependence.
    Used as the null in permutation tests (Sections 5.1 and 5.5). The paper's own random baseline estimate is J≈0.01-0.05, so 0.5 is an arbitrary and generous null.
  • ad hoc to paper 1-J measures the proportion of scene description that is context-dependent, and 'context explains ~90% of variance'.
    Set overlap is not a variance decomposition; this equivalence is asserted in Section 5.6 without derivation.
  • domain assumption all-MiniLM-L6-v2 sentence embeddings capture semantic similarity between affordance descriptions.
    Used for cosine similarity and Tucker decomposition (Sections 4.2 and 5.8); no validation on this output domain.
  • domain assumption COCO-2017 validation images selected for 'multi-object scenes with high interaction potential' are representative of naturalistic scenes for affordance generalization.
    Corpus selection described in Section 4.1; no diversity analysis beyond COCO categories.

pith-pipeline@v1.3.0-alltime-deepseek · 19963 in / 13927 out tokens · 127886 ms · 2026-08-02T23:28:39.526615+00:00 · methodology

0 comments
read the original abstract

We characterize the phenomenon of context-dependent affordance computation in vision-language models (VLMs). Our primary study uses Qwen3-VL-30B-A3B ($n = 3{,}213$ scene-context pairs from COCO-2017: 479 images under 7 agentic personas), with a cross-model replication on LLaVA-1.5-13B. We demonstrate substantial affordance drift: mean Jaccard similarity between context conditions is $0.095$ (95% CI $[0.092, 0.097]$ across $N = 479$ images; $9{,}244$ prime pairs; $p < 0.0001$), indicating that more than 90% of lexical scene description is context-dependent; the LLaVA replication reproduces the effect (mean $J = 0.160$, 84% context-dependent). Sentence-level cosine similarity confirms drift at the semantic level (mean $= 0.415$, 58.5% context-dependent). Stochastic baseline experiments ($2{,}384$ inference runs across 4 temperatures and 5 seeds) confirm this reflects genuine context effects rather than generation noise: within-prime variance is substantially lower than cross-prime variance across all conditions. Tucker decomposition with bootstrap stability analysis ($n = 1{,}000$ resamples) reveals stable orthogonal latent factors: a "Culinary Manifold" isolated to chef contexts and an "Access Axis" spanning child-mobility contrasts. The gap between lexical (90%) and semantic (58.5%) measures indicates that surface vocabulary changes more than underlying meaning under context shifts. These findings suggest a direction for robotics: dynamic, query-dependent ontological projection (JIT Ontology) rather than static world modeling. We do not claim to establish processing order or architectural primacy; such claims require internal representational analysis beyond output behavior.

Figures

Figures reproduced from arXiv: 2603.04419 by Murad Farzulla.

Figure 1
Figure 1. Figure 1: Comparison of visual processing pipelines. (a) Standard computer vision computes geometry [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Distribution of pairwise Jaccard similarity between context primes ( [PITH_FULL_IMAGE:figures/full_fig_p014_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Tucker decomposition factor loadings for context primes. Dim [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

44 extracted references · 9 canonical work pages

  1. [1]

    Do as I can, not as I say: Grounding language in robotic affordances

    Michael Ahn, Anthony Brohan, Noah Brown, Yevgen Chebotar, Omar Cortes, Byron David, Chelsea Finn, et al. Do as I can, not as I say: Grounding language in robotic affordances. In Conference on Robot Learning (CoRL). PMLR, 2022

    2022

  2. [2]

    Qwen-VL : A versatile vision-language model for understanding, localization, text reading, and beyond

    Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. Qwen-VL : A versatile vision-language model for understanding, localization, text reading, and beyond. arXiv preprint arXiv:2308.12966, 2023

    Pith/arXiv arXiv 2023

  3. [3]

    Bender and Alexander Koller

    Emily M. Bender and Alexander Koller. Climbing towards NLU : On meaning, form, and understanding in the age of data. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics (ACL), pages 5185--5198. ACL, 2020. doi:10.18653/v1/2020.acl-main.463

    doi:10.18653/v1/2020.acl-main.463 2020

  4. [4]

    Andrei Z. Broder. On the resemblance and containment of documents. In Compression and Complexity of Sequences, pages 21--29. IEEE, 1997. doi:10.1109/SEQUEN.1997.666900

    arXiv 1997

  5. [5]

    An outline of a theory of affordances

    Anthony Chemero. An outline of a theory of affordances. Ecological Psychology, 15 0 (2): 0 181--195, 2003. doi:10.1207/s15326969eco1502_5

    doi:10.1207/s15326969eco1502_5 2003

  6. [6]

    Cortical mechanisms of action selection: The affordance competition hypothesis

    Paul Cisek. Cortical mechanisms of action selection: The affordance competition hypothesis. Philosophical Transactions of the Royal Society B, 362 0 (1485): 0 1585--1599, 2007. doi:10.1098/rstb.2007.2054

    arXiv 2007

  7. [7]

    Being There: Putting Brain, Body, and World Together Again

    Andy Clark. Being There: Putting Brain, Body, and World Together Again. MIT Press, 1997

    1997

  8. [8]

    Whatever next? P redictive brains, situated agents, and the future of cognitive science

    Andy Clark. Whatever next? P redictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36 0 (3): 0 181--204, 2013. doi:10.1017/S0140525X12000477

    doi:10.1017/s0140525x12000477 2013

  9. [9]

    Hubert L. Dreyfus. What Computers Still Can't Do: A Critique of Artificial Reason. MIT Press, 1992

    1992

  10. [10]

    Hubert L. Dreyfus. Why H eideggerian AI failed and how fixing it would require making it more H eideggerian. Philosophical Psychology, 20 0 (2): 0 247--268, 2007. doi:10.1080/09515080701239510

    doi:10.1080/09515080701239510 2007

  11. [11]

    Training data and the maladaptive mind: A computational framework for developmental trauma

    Murad Farzulla. Training data and the maladaptive mind: A computational framework for developmental trauma. Research Square, 2025. doi:10.21203/rs.3.rs-8634152/v1. Under review at Humanities & Social Sciences Communications (Nature). Zenodo: 10.5281/zenodo.17681336

    work page doi:10.21203/rs.3.rs-8634152/v1 2025

  12. [12]

    James J. Gibson. The Senses Considered as Perceptual Systems. Houghton Mifflin, 1966

    1966

  13. [13]

    James J. Gibson. The Ecological Approach to Visual Perception. Houghton Mifflin, 1979

    1979

  14. [14]

    Towards robust visual understanding: A paradigm shift in computer vision from recognition to reasoning

    Tejas Gokhale. Towards robust visual understanding: A paradigm shift in computer vision from recognition to reasoning. AI Magazine, 45 0 (3): 0 396--403, 2024. doi:10.1002/aaai.12194. AAAI New Faculty Highlights invited talk

    work page doi:10.1002/aaai.12194 2024

  15. [15]

    Goodale and A

    Melvyn A. Goodale and A. David Milner. Separate visual pathways for perception and action. Trends in Neurosciences, 15 0 (1): 0 20--25, 1992. doi:10.1016/0166-2236(92)90344-8

    doi:10.1016/0166-2236(92)90344-8 1992

  16. [16]

    Visual affordance and function understanding: A survey

    Mohammed Hassanin, Salman Khan, and Murat Tahtali. Visual affordance and function understanding: A survey. ACM Computing Surveys, 54 0 (3): 0 1--35, 2022. doi:10.1145/3446370

    work page doi:10.1145/3446370 2022

  17. [17]

    Sein und Zeit

    Martin Heidegger. Sein und Zeit. Max Niemeyer Verlag, 1927

    1927

  18. [18]

    Leverage task context for object affordance ranking

    Haojie Huang, Hongchen Luo, Wei Zhai, Yang Cao, and Zheng-Jun Zha. Leverage task context for object affordance ranking. arXiv preprint arXiv:2411.16082, 2024

    Pith/arXiv arXiv 2024

  19. [19]

    VoxPoser : Composable 3D value maps for robotic manipulation with language models

    Wenlong Huang, Chen Wang, Ruohan Zhang, Yunzhu Li, Jiajun Wu, and Li Fei-Fei. VoxPoser : Composable 3D value maps for robotic manipulation with language models. In Conference on Robot Learning (CoRL). PMLR, 2023

    2023

  20. [20]

    Shamma, et al

    Ranjay Krishna, Yuke Zhu, Oliver Groth, Justin Johnson, Kenji Hata, Joshua Kravitz, Stephanie Chen, Yannis Kalantidis, Li-Jia Li, David A. Shamma, et al. Visual genome: Connecting language and vision using crowdsourced dense image annotations. International Journal of Computer Vision, 123 0 (1): 0 32--73, 2017. doi:10.1007/s11263-016-0981-7

    doi:10.1007/s11263-016-0981-7 2017

  21. [21]

    Lawrence Zitnick

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll \'a r, and C. Lawrence Zitnick. Microsoft COCO : Common objects in context. In European Conference on Computer Vision (ECCV), pages 740--755. Springer, 2014. doi:10.1007/978-3-319-10602-1_48

    doi:10.1007/978-3-319-10602-1_48 2014

  22. [22]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. In Advances in Neural Information Processing Systems (NeurIPS), 2024

    2024

  23. [23]

    Urbano Lorenzo-Seva and Jos M. F. ten Berge. Tucker's congruence coefficient as a meaningful index of factor similarity. Methodology, 2 0 (2): 0 57--64, 2006. doi:10.1027/1614-2241.2.2.57

    doi:10.1027/1614-2241.2.2.57 2006

  24. [24]

    Ph \'e nom \'e nologie de la Perception

    Maurice Merleau-Ponty. Ph \'e nom \'e nologie de la Perception . Gallimard, 1945

    1945

  25. [25]

    Object representation in the ventral premotor cortex (area F5 ) of the monkey

    Akira Murata, Luciano Fadiga, Leonardo Fogassi, Vittorio Gallese, Vassilis Raos, and Giacomo Rizzolatti. Object representation in the ventral premotor cortex (area F5 ) of the monkey. Journal of Neurophysiology, 78 0 (4): 0 2226--2230, 1997. doi:10.1152/jn.1997.78.4.2226

    work page doi:10.1152/jn.1997.78.4.2226 1997

  26. [26]

    RT-Affordance : Affordances are versatile intermediate representations for robot manipulation

    Soroush Nasiriany, Sean Kirmani, Tianhe Ding, Laura Smith, Yuke Zhu, Danny Driess, Dorsa Sadigh, and Ted Xiao. RT-Affordance : Affordances are versatile intermediate representations for robot manipulation. arXiv preprint arXiv:2411.02704, 2024. doi:10.1109/icra55743.2025.11127525

    Pith/arXiv arXiv 2024

  27. [27]

    Caldwell, and Nikos G

    Anh Nguyen, Dimitrios Kanoulas, Darwin G. Caldwell, and Nikos G. Tsagarakis. Object-based affordances detection with convolutional neural networks and dense conditional random fields. In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 5908--5915, 2017. doi:10.1109/IROS.2017.8206484

    arXiv 2017

  28. [28]

    Action in Perception

    Alva No \"e . Action in Perception. MIT Press, 2004

    2004

  29. [29]

    Kevin O'Regan and Alva No \"e

    J. Kevin O'Regan and Alva No \"e . A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences, 24 0 (5): 0 939--973, 2001. doi:10.1017/S0140525X01000115

    doi:10.1017/s0140525x01000115 2001

  30. [30]

    AffordanceLLM : Grounding affordance from vision language models

    Shengyi Qian, Weifeng Chen, Min Bai, Xiong Zhou, Zhuowen Tu, and Li Erran Li. AffordanceLLM : Grounding affordance from vision language models. In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, pages 7587--7597, 2024. doi:10.1109/CVPRW63382.2024.00754

    arXiv 2024

  31. [31]

    Sentence- BERT : Sentence embeddings using S iamese BERT -networks

    Nils Reimers and Iryna Gurevych. Sentence- BERT : Sentence embeddings using S iamese BERT -networks. In Conference on Empirical Methods in Natural Language Processing (EMNLP-IJCNLP), pages 3982--3992. ACL, 2019. doi:10.18653/v1/D19-1410

    doi:10.18653/v1/d19-1410 2019

  32. [32]

    The mirror-neuron system

    Giacomo Rizzolatti and Laila Craighero. The mirror-neuron system. Annual Review of Neuroscience, 27: 0 169--192, 2004. doi:10.1146/annurev.neuro.27.070203.144230

    arXiv 2004

  33. [33]

    A survey on efficient vision-language models

    Gaurav Shinde, Anuradha Ravi, Emon Dey, Shadman Sakib, Milind Rampure, and Nirmalya Roy. A survey on efficient vision-language models. WIREs Data Mining and Knowledge Discovery, 15 0 (3): 0 e70036, 2025. doi:10.1002/widm.70036

    work page doi:10.1002/widm.70036 2025

  34. [34]

    Simons and Christopher F

    Daniel J. Simons and Christopher F. Chabris. Gorillas in our midst: Sustained inattentional blindness for dynamic events. Perception, 28 0 (9): 0 1059--1074, 1999. doi:10.1068/p281059

    doi:10.1068/p281059 1999

  35. [35]

    Stoffregen

    Thomas A. Stoffregen. Affordances as properties of the animal-environment system. Ecological Psychology, 15 0 (2): 0 115--134, 2003. doi:10.1207/s15326969eco1502_2

    work page doi:10.1207/s15326969eco1502_2 2003

  36. [36]

    Differential effects of contextual congruency on recognition and retrieval of perceptual details

    Rebeca Su \'a rez, Bruno Lara, and Alejandra Ciria. Differential effects of contextual congruency on recognition and retrieval of perceptual details. Memory & Cognition, 2026. doi:10.3758/s13421-025-01848-0. Published online 3 February 2026

    work page doi:10.3758/s13421-025-01848-0 2026

  37. [37]

    Mind in Life: Biology, Phenomenology, and the Sciences of Mind

    Evan Thompson. Mind in Life: Biology, Phenomenology, and the Sciences of Mind. Harvard University Press, 2007. ISBN 978-0-674-02511-0

    2007

  38. [38]

    Ledyard R. Tucker. Some mathematical notes on three-mode factor analysis. Psychometrika, 31 0 (3): 0 279--311, 1966. doi:10.1007/BF02289464

    doi:10.1007/bf02289464 1966

  39. [39]

    Michael T. Turvey. Affordances and prospective control: An outline of the ontology. Ecological Psychology, 4 0 (3): 0 173--187, 1992. doi:10.1207/s15326969eco0403_3

    work page doi:10.1207/s15326969eco0403_3 1992

  40. [40]

    Varela, Evan Thompson, and Eleanor Rosch

    Francisco J. Varela, Evan Thompson, and Eleanor Rosch. The Embodied Mind: Cognitive Science and Human Experience. MIT Press, 1991

    1991

  41. [41]

    Choy, and Li Fei-Fei

    Danfei Xu, Yuke Zhu, Christopher B. Choy, and Li Fei-Fei. Scene graph generation by iterative message passing. In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 5410--5419, 2017. doi:10.1109/CVPR.2017.330

    work page doi:10.1109/cvpr.2017.330 2017

  42. [42]

    RoboPoint : A vision-language model for spatial affordance prediction for robotics

    Wentao Yuan, Jiafei Duan, Valts Blukis, Wilbert Pumacay, Ranjay Krishna, Adithyavairavan Murali, Arsalan Mousavian, and Dieter Fox. RoboPoint : A vision-language model for spatial affordance prediction for robotics. In Conference on Robot Learning (CoRL), pages 4005--4020. PMLR, 2024

    2024

  43. [43]

    Self-explainable affordance learning with embodied caption

    Zhipeng Zhang, Zhimin Wei, Guolei Sun, Peng Wang, and Luc Van Gool. Self-explainable affordance learning with embodied caption. arXiv preprint arXiv:2404.05603, 2024

    Pith/arXiv arXiv 2024

  44. [44]

    Places: A 10 million image database for scene recognition

    Bolei Zhou, Agata Lapedriza, Aditya Khosla, Aude Oliva, and Antonio Torralba. Places: A 10 million image database for scene recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40 0 (6): 0 1452--1464, 2017. doi:10.1109/TPAMI.2017.2723009

    arXiv 2017