Pith. sign in

REVIEW 4 major objections 5 minor 41 references

Hybrid Reasoning for Perception, Explanation, and Autonomous Action in Manufacturing

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Embedding a convolutional regression expert into a vision-language model gives a 3D printer engineering-grade flow-rate perception and lets it turn text or image prompts into executable G-code.

desk verdict Solid hybrid-VLA architecture with a convincing firmware-grounded regression result, but 'strong OOD generalization' is not established by the LLM-in-the-loop evaluation. read the letter →

arxiv 2506.08462 v1 pith:YFDFYJ4M submitted 2025-06-10 cs.AI cs.HCcs.ROcs.SYeess.SY

Christos Margadji , Sebastian W. Pattinson This is my paper
classification cs.AIcs.HCcs.ROcs.SYeess.SY
keywords vision-language-actionmodelshybridreasoning3Dprintingcontrolprocessexpertregressioninfoundationretrieval-augmentedgenerationchain-of-thoughtautonomousG-code
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 introduces CIPHER, a vision-language-action framework for industrial control, tested on a commercial 3D printer. It argues that pure vision-language models cannot perform the continuous regression that engineering perception requires, because next-token classification and stochastic generation fail to converge on numeric values. Embedding a small convolutional regression expert that maps nozzle images to a flow-rate token and feeds it into the language decoder reduces mean absolute error roughly fivefold and makes qualitative descriptions consistent with the numbers. With retrieval-augmented domain knowledge and a G-code codebook, the same agent answers physics-informed questions, explains its reasoning, and generates executable machine instructions; a geometry module lets it print from text or image prompts without manual CAD. The paper's claim is that modular hybrid reasoning of this kind gives foundation models the precision and transparency needed for data-scarce manufacturing.

What carries the argument

The load-bearing mechanism is the process expert: a 116-million-parameter residual convolutional network that takes a 224-by-224 nozzle image, normalizes it with the pretraining dataset's statistics, regresses the flow rate in logarithmic space, and is then projected to a single token that the autoregressive language decoder conditions on under a causal mask. Because the numeric value enters as a token rather than being reconstructed by next-token classification, the continuous state is available to the decoder reliably. Around this core, the framework adds three supporting modules: a retrieval-augmented generation module that retrieves the five nearest facts by cosine similarity from a pre-embedded 3,930-fact knowledge base; a physics expert that turns observations into chain-of-thought explanations and G-code commands, using a parsed G-code playbook from the printer firmware; and a geometry expert that routes simple shape requests to primitive-composition Python code and complex requests to a pre-trained text-to-3D and image-to-3D shape generator, validating output through a slicer engine.

What would settle it

Have a certified process engineer score the same 100 out-of-distribution reasoning challenges that the machine judge scored, then compare the two rankings; if the human ordering does not reproduce the paper's ranking, or if physically executing the agent's recommended corrections does not improve print quality, the generalization claim is falsified.

Watch

Extended reading notes

Core claim

The central claim is that a vision-language-action model built by attaching a convolutional regression network to a multimodal language model can perceive continuous physical states and act on them, where the base vision-language model alone cannot. In their instantiation, a residual convolutional 'process expert' reads endoscope images of the nozzle, predicts extrusion flow rate in log space, and injects the prediction into the token stream as a conditioning start token. On 1,000 held-out samples this yields a mean absolute error of 17.52 with standard deviation 28.89, versus 82.92 with standard deviation 71.83 without the expert, with classification accuracy rising from about 30 percent to about 88 percent; the gain persists even when only the projection layer is trained, which also preserves most pre-trained language and vision knowledge. The paper further claims that retrieval-augmented generation from a 3,930-fact domain knowledge base improves question answering and physics-informed chain-of-thought reasoning enough to reach pairwise ranking scores comparable to a frontier model, and that the agent can translate natural-language or image prompts into G-code and print simple and complex shapes one-shot. On the paper's terms, the process expert overcomes a structural limitation of vision-language models, and the modular design makes precise, explainable, autonomous manufacture possible without explicit annotations.

Load-bearing premise

The reasoning and out-of-distribution claims assume that the machine-generated facts in the knowledge base and the machine-based judge used to score answers are actually correct; if those facts or judgments are wrong, the claimed gains in reasoning and generalization collapse, even though the flow-rate regression itself is grounded in the printer's firmware labels.

Editorial extensions

If this is right

  • If the central claim is right, any continuous physical quantity in manufacturing, such as temperature, pressure, feed rate, or dimensional deviation, could be handled by the same recipe: a small task-specific regressor feeding tokens into a frozen or lightly adapted language model.
  • The preferred configuration, which trains only a projection layer, shows that domain-specific quantitative perception can be added to a foundation model while preserving almost all prior language and vision skills and avoiding catastrophic forgetting.
  • Retrieval-augmented generation plus a structured G-code codebook raises chain-of-thought decision quality to a level comparable to a frontier model, even though the base model is far smaller.
  • Prompt-to-print geometry generation works one-shot for primitive compositions and many complex shapes, with failures concentrated in support structures and minimum feature sizes that the shape generator does not model.
  • Low-rank adaptation cuts memory consumption by 52.4 percent and the volume of training data required by 81 percent, while slightly reducing quantitative and qualitative accuracy relative to full fine-tuning with the process expert.

Reading between the lines

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

  • Editorial extension: if the process-expert token-injection mechanism is the cause of the gain, the same architecture should improve regression of other continuous quantities, such as nozzle temperature or adhesion force, with only the expert head and projection layer retrained; this is directly testable on the same printer.
  • Editorial extension: the evaluation loop uses machine-generated facts and a machine-based judge, so the reasoning results should be re-scored by human engineers before the claimed out-of-distribution generalization is treated as established.
  • Editorial extension: the growing library of validated geometry functions is a form of self-improving procedural memory; one could test whether printing success rates keep rising across a sequence of novel prompts without any additional training.
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

4 major / 5 minor

Summary. The paper introduces CIPHER, a hybrid vision-language-action (VLA) architecture for industrial control, instantiated in a 3D printer. The system couples a Llama-3.2 vision-language backbone with a ResNet-152 'process expert' that performs quantitative regression on nozzle images, a retrieval-augmented generation (RAG) module built from LLM-generated facts, and a 'geometry expert' that routes shape-generation requests to either primitive-composition code (generated via GPT-4o) or Shap-E. The central claims are that the process expert yields a five-fold reduction in flow-rate MAE (from 82.92 to 17.52), that LoRA-based fine-tuning preserves prior knowledge, that RAG enables physics-informed chain-of-thought reasoning and 'strong generalization to out-of-distribution tasks,' and that the full pipeline can autonomously produce G-code and even print from text or image prompts.

Significance. If the claims were fully substantiated, CIPHER would be a valuable example of combining a small, task-specific regression module with a large foundation model to achieve quantitative perception and interpretable control in a data-scarce manufacturing setting. The process-expert result is grounded in firmware labels, and the authors report detailed hyperparameters, a code repository, and some physical printing demonstrations, which are concrete strengths that aid reproducibility. However, the advertised out-of-distribution reasoning and explanation gains rest on an evaluation loop in which the knowledge base, the test questions, and the judge are all produced by large language models, with no human or physical ground truth; this is a load-bearing weakness for the paper's broadest claims. The G-code control experiment is also performed on synthetically constructed inputs rather than on the physical printer, further limiting the strength of the 'precise machine instructions' claim.

major comments (4)
  1. [Section 2.1, Table A1] The reported five-fold MAE reduction is internally inconsistent. The text states 'five-fold reduction in MAE when comparing results from E1 (=82.92±71.83) and E5 (=17.62±30.04)', but Table A1 lists E5 as 17.78±29.95 and E4 as 17.62±30.04, with E2 at 17.52±28.89. The authors must clarify which configuration corresponds to the reported value and why Section 2.2 later compares against E5 as the strongest baseline. Furthermore, the differences among E2, E3, E4, and E5 are all within one standard deviation of one another; a significance test or confidence interval is needed before claiming that the process expert 'effectively overcomes limitations inherent to the base architecture' for the E2 configuration.
  2. [Sections 2.4.2, 4.1.2, 4.3.5] The out-of-distribution reasoning and explanation claims are evaluated entirely inside an LLM-generated loop. The RAG knowledge base is 'populated by prompting an LLM to generate factual statements' (4.1.2), the test questions are derived from a mind map 'constructed with the assistance of an LLM' (2.4.2), and the Elo ratings are assigned by a GPT-4o instance (4.3.5). No human or physical ground truth checks the correctness of the answers, and the reported Elo differences (e.g., Ours+codebook 1238±9 vs. GPT-4o 1241±11 in Table A3) overlap substantially. The abstract's headline claim of 'strong generalization to out-of-distribution tasks' is therefore not established; the results could reflect the judge's preference for longer or more structured text rather than engineering correctness. The authors should temper this claim or provide a human-evaluated subset and a physical verification of recommended actions (e.g., printing under the suggested 240°C ABS setting).
  3. [Section 2.4.1] The 'performing traditional control' experiment does not actually close the loop on the printer. The 100 (estimate, firmware) pairs are synthesized by adding noise within the empirical error bounds around process-expert estimates, and the models only produce text containing an M221 command. The reported 'additional control error of 0.215±0.090' is the error of the generated numeric command relative to the target adjustment, which trivially measures arithmetic formatting rather than physical control performance. Without executing the commands on the printer or in a realistic simulator that couples the resulting flow rate to actual extrusion, the claim that the system 'autonomously generates precise machine instructions' is not demonstrated.
  4. [Sections 2.5, 4.2.3] The geometry expert's contributions are not cleanly separated from external tools, and the success criteria are unspecified. The complexity routing, Python code generation for primitives, and the text-to-shape branch use GPT-4o and Shap-E APIs; the paper does not state how much of the reported 'autonomous fabrication' is due to CIPHER's own reasoning versus these external components. Moreover, 'all geometric primitive requests succeeded in a one-shot manner' is asserted without a quantitative definition of success (e.g., dimensional accuracy, print completion, or visual inspection criteria), and several complex shapes (llama legs, sunglass temples) are acknowledged to be unprintable. The authors should provide measurable success rates and clarify CIPHER's specific contribution to the geometry pipeline.
minor comments (5)
  1. [Throughout] There are several typographical errors: 'T able A1' and 'T able A2' in the appendix, 'RGA' instead of 'RAG' in Table A2, 'resoning' in the Fig. 5c caption, and 'extend by which' in Section 2.3. The manuscript would benefit from a careful proofreading pass.
  2. [Methods 4.1.2, Appendix figures] The mind-map citation is inconsistent: Section 4.1.2 refers to 'Extended Data Fig. A4' for the mind map, while Section 2.4.2 and the appendix label the mind map as Fig. A3. Please align the cross-references.
  3. [Equation (5)] The Elo expected-score formula is unclear as typeset: 'E = 1 / (1 + 10 × Ropp − R / 400)' should be E = 1 / (1 + 10^((R_opp − R)/400)). Please correct the notation.
  4. [Section 4.2.2] The training details for the process expert are incomplete: the text mentions a linear learning-rate scheduler with γ=0.5 and 'original value =1e-4', but Extended Data Table A4 lists warmup iterations as 500,000, which seems inconsistent with a single epoch of 20 hours on the described dataset. Please clarify the actual learning-rate schedule and the number of warmup steps.
  5. [Section 2.4.1] The command notation 'M221 SN' should likely be 'M221 S<N>' or similar; please use standard G-code syntax so the reader can verify the intended command.

Circularity Check

3 steps flagged · score 6.0 of 10

Flow-rate regression is firmware-grounded and independent, but the OOD knowledge, chain-of-thought, and control-error claims are evaluated with LLM-generated facts and an LLM judge, so those predictions reduce to self-consistency.

  1. self definitional [Section 2.4.2 'Domain knowledge and knowledge injection'; Methods 4.1.2 'Knowledge for retrieval-augmented generation'; Methods 4.3.5 'Elo ranking']
    "This mind-map was constructed with the assistance of an LLM to ensure comprehensive topic coverage. / ... which itself was populated by prompting an LLM to generate factual statements for each node of the said mind-map. / To determine scores for a pair of answers, we consulted a GPT-4o instance using Supplementary Material Prompt S3 for question-answering and Supplementary Material Prompt S4 for physics-informed chain-of-thought reasoning."

    The reported 'net gain in factual accuracy' and Elo superiority (Ours+RGA 1238 vs GPT-4o 1241, Table A2) are not measured against any human or physical ground truth. Equation (4)'s score S is assigned by a GPT-4o instance, and the RAG knowledge repository is 'populated by prompting an LLM to generate factual statements'. The questions themselves are organized by an LLM-built mind map. The claimed factual-accuracy improvement therefore reduces, by construction, to GPT-4o judging text that was generated with GPT-4o-derived facts; it measures self-consistency, not external correctness.

  2. other [Section 2.4.2 'Emerging behaviors'; Methods 4.3.5 'Elo ranking']
    "First we validate the chain-of-thought reasoning quality using the Elo ranking system like before, with results summarized in Figure 5c. / In each trial, two models were randomly selected, and their respective answers were blindly evaluated by a third-party reviewer (implemented as an LLM), which judged responses based on correctness and contextual relevance."

    The emerging-behaviors benchmark has no known-optimal action: proposed control commands are never executed, and 'chain-of-thought reasoning quality' is scored by Elo with a GPT-4o judge. The abstract's headline 'strong generalization to out-of-distribution tasks' is thus supported by a closed loop in which the same LLM family generates the knowledge base and assigns the correctness scores. No independent physical or human check verifies that the recommended corrective strategies (e.g., '240°C for ABS') are safe or optimal, so the OOD reasoning 'prediction' is an LLM-preference score by construction.

1 more flagged steps
  1. fitted input called prediction [Section 2.4.1 'Performing traditional control']
    "we synthetically derive synthetic firmware values that represent the printer’s intrinsic measurement, or its belief, by sampling within a range defined by the said empirical error bounds, centered around the corresponding{estimate}."

    The reported 'marginal additional control error of 0.215±0.090' is computed against a firmware 'belief' that was itself synthesized from the process expert's own estimate and its empirical error bounds. The agent is therefore never tested against an independently measured physical flow state; the control-precision figure is, by construction, a self-consistency check between the process expert's output distribution and the command-generation arithmetic, not an independent measurement of physical control error.

full rationale

The flow-rate regression (Section 2.1) is grounded in firmware labels and evaluated on a held-out test set (MAE 17.62±30.04 vs 82.92±71.83), so that part of the derivation is independent and not circular. The LoRA efficiency numbers and the physically printed geometry results (Section 2.5) are also externally grounded. However, the knowledge-reasoning and chain-of-thought 'predictions' are evaluated inside an LLM-generated loop: the RAG facts are generated by prompting an LLM, the mind-map/test questions are LLM-assisted, and the Elo judge is GPT-4o. The reported factual-accuracy and reasoning-quality gains therefore reduce, by construction, to GPT-4o's pairwise preference over text partly generated from GPT-4o-derived facts. The control experiment similarly synthesizes the firmware 'belief' from the process expert's own estimate and error bounds, so the reported control error is a self-consistency measurement rather than an independent physical control result. The paper itself concedes explanation reliability 'warrants further scrutiny' and that performance 'remains contingent on the quality of its expert modules and knowledge retrieval mechanisms', which is consistent with this partial circularity. No load-bearing self-citation chain or imported uniqueness theorem was found.

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

The framework introduces no new physical entities; the free parameters are hand-chosen thresholds and retrieval counts that affect the reported metrics, and the axioms are domain assumptions about label quality, LLM-generated knowledge, LLM-based evaluation, and the synthetic captioning scheme. The flow-rate regression itself has independent firmware labels, which limits the circularity burden of that specific result.

free parameters (4)
  • Flow-rate qualitative thresholds = 90% and 110%
    Used to bin raw flow values into under/good/over-extrusion classes that define the caption templates and the classification-accuracy metric; chosen by hand in Methods 4.1.2.
  • RAG retrieval count = N=5
    Number of nearest knowledge facts injected into prompts at inference; chosen in Methods 4.1.2 without apparent tuning rationale.
  • Geometry complexity threshold = N=4 primitives
    Requests needing more than four geometric primitives are routed to the Shap-E branch instead of primitive composition; set in Methods 4.2.3.
  • Log-space flow mapping bounds = 30-300 to approx -1 to 1
    Output normalization for the process expert; a design choice that affects regression scaling and could influence error distribution.
assumptions (4)
  • domain assumption Firmware flow-rate values are reliable ground truth for extrusion quality.
    Used to generate all training labels in Methods 4.1; if firmware reporting is inaccurate, the process-expert target is corrupted.
  • domain assumption The LLM-generated knowledge base contains correct, non-hallucinated domain facts.
    RAG facts are produced by prompting an LLM from an LLM-built mind map (Methods 4.1.2); no human verification is reported.
  • domain assumption GPT-4o as judge provides a valid, unbiased evaluation of reasoning and answer quality.
    Elo ratings in Section 2.4.2 and Methods 4.3.5 are assigned by a GPT-4o instance using LLM prompts; no human inter-rater agreement is reported.
  • ad hoc to paper Synthetic captions preserve the quantitative information needed to learn regression.
    Captions are assembled from templates (general + quantitative + qualitative) in Methods 4.1.2; the regression signal is only the embedded numeric phrase, and template artifacts could be exploited.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid Reasoning for Perception, Explanation, and Autonomous Action in Manufacturing." pith.science (2026). https://pith.science/paper/YFDFYJ4M

@misc{pith2026250608462,
  author       = {Pith},
  title        = {Pith review of: Hybrid Reasoning for Perception, Explanation, and Autonomous Action in Manufacturing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/YFDFYJ4M}},
  note         = {Machine review of arXiv:2506.08462}
}
read the original abstract

Industrial processes must be robust and adaptable, as environments and tasks are often unpredictable, while operational errors remain costly and difficult to detect. AI-based control systems offer a path forward, yet typically depend on supervised learning with extensive labelled datasets, which limits their ability to generalize across variable and data-scarce industrial settings. Foundation models could enable broader reasoning and knowledge integration, but rarely deliver the quantitative precision demanded by engineering applications. Here, we introduceControl and Interpretation of Production via Hybrid Expertise and Reasoning (CIPHER): a vision-language-action (VLA) model framework aiming to replicate human-like reasoning for industrial control, instantiated in a commercial-grade 3D printer. It integrates a process expert, a regression model enabling quantitative characterization of system states required for engineering tasks. CIPHER also incorporates retrieval-augmented generation to access external expert knowledge and support physics-informed, chain-of-thought reasoning. This hybrid architecture exhibits strong generalization to out-of-distribution tasks. It interprets visual or textual inputs from process monitoring, explains its decisions, and autonomously generates precise machine instructions, without requiring explicit annotations. CIPHER thus lays the foundations for autonomous systems that act with precision, reason with context, and communicate decisions transparently, supporting safe and trusted deployment in industrial settings.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

41 extracted references · 25 canonical work pages

  1. [1]

    (2023).https://api.semanticscholar.org/CorpusID:257532815 29

    Achiam, O.J., Adler, S., Agarwal, S., Ahmad, L., Akkaya, I., Aleman, F.L., Almeida, D., Altenschmidt, J., Altman, S., Anadkat, S., Avila, R., Babuschkin, I., Balaji, S., Balcom, V., Baltescu, P., Bao, H.-i., Bavarian, M., Belgum, J., Bello, I., Berdine, J., Bernadett-Shapiro, G., Berner, C., Bogdonoff, L., Boiko, O., Boyd, M.-l., Brakman, A.-L., Brockman,...

    work page 2023

  2. [2]

    ArXiv abs/2302.13971(2023)

    Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.-A., Lacroix, T., Rozi` ere, B., Goyal, N., Hambro, E., Azhar, F., Rodriguez, A., Joulin, A., Grave, E., Lample, G.: Llama: Open and efficient foundation language models. ArXiv abs/2302.13971(2023)

    arXiv 2023

  3. [3]

    In: Oh, A., Naumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S

    Liu, H., Li, C., Wu, Q., Lee, Y.J.: Visual instruction tuning. In: Oh, A., Naumann, T., Globerson, A., Saenko, K., Hardt, M., Levine, S. (eds.) Advances in Neural Information Processing Systems, vol. 36, pp. 34892–34916. Curran Associates, Inc., ??? (2023)

    work page 2023

  4. [4]

    In: Tan, J., Toussaint, M., Darvish, K

    Zitkovich, B., Yu, T., Xu, S., Xu, P., Xiao, T., Xia, F., Wu, J., Wohlhart, P., Welker, S., Wahid, A., Vuong, Q., Vanhoucke, V., Tran, H., Soricut, R., Singh, A., Singh, J., Sermanet, P., Sanketi, P.R., Salazar, G., Ryoo, M.S., Reymann, K., Rao, K., Pertsch, K., Mordatch, I., Michalewski, H., Lu, Y., Levine, S., Lee, L., Lee, T.-W.E., Leal, I., Kuang, Y.,...

    work page 2023

  5. [5]

    ArXivabs/2205.06175(2022)

    Reed, S., Zolna, K., Parisotto, E., Colmenarejo, S.G., Novikov, A., Barth-Maron, G., Gim´ enez, M., Sulsky, Y., Kay, J., Springenberg, J.T., Eccles, T., Bruce, J., Razavi, A., Edwards, A.D., Heess, N.M.O., Chen, Y., Hadsell, R., Vinyals, O., Bordbar, M., Freitas, N.: A generalist agent. ArXivabs/2205.06175(2022)

    arXiv 2022

  6. [6]

    ArXivabs/2401.10463(2024)

    Zhu, X., Fu, Y., Zhou, B., Lin, Z.: Critical data size of language models from a grokking perspective. ArXivabs/2401.10463(2024)

    arXiv 2024

  7. [7]

    In: Proceedings of the 37th International Conference on Neural Information Processing Systems

    Muennighoff, N., Rush, A.M., Barak, B., Scao, T.L., Piktus, A., Tazi, N., Pyysalo, S., Wolf, T., Raffel, C.: Scaling data-constrained language models. In: Proceedings of the 37th International Conference on Neural Information Processing Systems. Curran Associates Inc., ??? (2024)

    work page 2024

  8. [8]

    Wang, G., Xie, Y., Jiang, Y., Mandlekar, A., Xiao, C., Zhu, Y., Fan, L.J., Anand- kumar, A.: Voyager: An open-ended embodied agent with large language models. Trans. Mach. Learn. Res.2024(2023)

    work page 2023

Show all 41 references
  1. [9]

    ArXivabs/2307.13854(2023)

    Zhou, S., Xu, F.F., Zhu, H., Zhou, X., Lo, R., Sridhar, A., Cheng, X., Bisk, Y., Fried, D., Alon, U., Neubig, G.: Webarena: A realistic web environment for building autonomous agents. ArXivabs/2307.13854(2023)

    2023 arXiv

  2. [10]

    ArXivabs/2409.12514(2024) 30

    Wen, J., Zhu, Y., Li, J., Zhu, M., Wu, K., Xu, Z., Liu, N., Cheng, R., Shen, C., Peng, Y., Feng, F., Tang, J.: Tinyvla: Towards fast, data-efficient vision-language- action models for robotic manipulation. ArXivabs/2409.12514(2024) 30

    2024 arXiv

  3. [11]

    IEEE Transactions on Intelligent Vehicles (2023)

    Zhou, X., Liu, M., Yurtsever, E., ˇZagar, B.L., Zimmer, W., Cao, H., Knoll, A.C.: Vision language models in autonomous driving: A survey and outlook. IEEE Transactions on Intelligent Vehicles (2023)

    2023

  4. [12]

    In: Agrawal, P., Kroemer, O., Burgard, W

    Kim, M.J., Pertsch, K., Karamcheti, S., Xiao, T., Balakrishna, A., Nair, S., Rafailov, R., Foster, E.P., Sanketi, P.R., Vuong, Q., Kollar, T., Burchfiel, B., Tedrake, R., Sadigh, D., Levine, S., Liang, P., Finn, C.: Openvla: An open-source vision-language-action model. In: Agr...

    2025

  5. [13]

    Nature Machine Intelligence7, 521 (2025) https://doi

    Robot planning with llms. Nature Machine Intelligence7, 521 (2025) https://doi. org/10.1038/s42256-025-01036-4

    2025 doi

  6. [14]

    Transactions on Machine Learning Research (2024)

    Song, X., Li, O., Lee, C., Yang, B., Peng, D., Perel, S., Chen, Y.: Omnipred: Lan- guage models as universal regressors. Transactions on Machine Learning Research (2024)

    2024

  7. [15]

    https://arxiv.org/abs/2502.19645

    Kim, M.J., Finn, C., Liang, P.: Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success (2025). https://arxiv.org/abs/2502.19645

    2025 arXiv

  8. [16]

    In: 2024 12th International Conference on Control, Mechatronics and Automation (ICCMA), pp

    Azarafza, M., Nayyeri, M., Steinmetz, C., Staab, S., Rettberg, A.: Hybrid rea- soning based on large language models for autonomous car driving. In: 2024 12th International Conference on Control, Mechatronics and Automation (ICCMA), pp. 14–22. IEEE, ??? (2024). https://doi.org...

    2024

  9. [17]

    IF AC Proceedings Volumes45(22), 90–95 (2012) https://doi.org/10.3182/20120905-3-HR-2030.00180

    G¨ unther, M., Hertzberg, J., Mansouri, M., Pecora, F., Saffiotti, A.: Hybrid reason- ing in perception: A case study. IF AC Proceedings Volumes45(22), 90–95 (2012) https://doi.org/10.3182/20120905-3-HR-2030.00180 . 10th IF AC Symposium on Robot Control

    2012

  10. [18]

    Advanced Engi- neering Materials20, 1700552 (2017) https://doi.org/10.1002/adem.201700552

    Klippstein, H., Sanchez, A., Hassanin, H., Zweiri, Y., Seneviratne, L.: Fused depo- sition modeling for unmanned aerial vehicles (uavs): A review. Advanced Engi- neering Materials20, 1700552 (2017) https://doi.org/10.1002/adem.201700552

    2017 doi

  11. [19]

    Nature609, 709–717 (2022) https://doi.org/10.1038/ s41586-022-04988-4

    Zhang, K., Chermprayong, P., Xiao, F., Tzoumanikas, D., Dams, B., Kay, S., Kocer, B.B., Burns, A., Orr, L., Alhinai, T., Choi, C., Darekar, D.D., Li, W., Hirschmann, S., Soana, V., Ngah, S.A., Grillot, C., Sareh, S., Choubey, A., Margheri, L., Pawar, V.M., Ball, R.J., Williams...

    2022

  12. [20]

    Advanced Materials Technolo- gies2, 1600084 (2017) https://doi.org/10.1002/admt.201600084 31

    Pattinson, S.W., Hart, A.J.: Additive manufacturing of cellulosic materials with robust mechanics and antimicrobial functionality. Advanced Materials Technolo- gies2, 1600084 (2017) https://doi.org/10.1002/admt.201600084 31

    2017 doi

  13. [21]

    Advanced Functional Materials29, 1901815 (2019) https://doi.org/10

    Pattinson, S.W., Huber, M.E., Kim, S., Lee, J., Grunsfeld, S., Roberts, R., Dreifus, G., Meier, C., Liu, L., Hogan, N., Hart, A.J.: Additive manufactur- ing of biomechanically tailored meshes for compliant wearable and implantable devices. Advanced Functional Materials29, 1901...

    2019

  14. [22]

    Additive Manufacturing25, 131–150 (2019) https: //doi.org/10.1016/j.addma.2018.10.028

    Ford, S., Minshall, T.: Invited review article: Where and how 3d printing is used in teaching and education. Additive Manufacturing25, 131–150 (2019) https: //doi.org/10.1016/j.addma.2018.10.028

    2019 doi

  15. [23]

    (eds.) Chapter 1 - Introduction to additive manufacturing, pp

    Zhang, X., Liou, F.: In: Pou, J., Riveiro, A., Davim, J.P. (eds.) Chapter 1 - Introduction to additive manufacturing, pp. 1–31. Else- vier, ??? (2021). https://doi.org/10.1016/B978-0-12-818411-0.00009-4 . https://www.sciencedirect.com/science/article/pii/B9780128184110000094

    2021 doi

  16. [24]

    Journal of Composites Science6(2022) https://doi.org/10.3390/jcs6070202

    Baechle-Clayton, M., Loos, E., Taheri, M., Taheri, H.: Failures and flaws in fused deposition modeling (fdm) additively manufactured polymers and composites. Journal of Composites Science6(2022) https://doi.org/10.3390/jcs6070202

    2022 doi

  17. [25]

    The International Journal of Advanced Manufacturing Technology 84, 1483–1495 (2016) https://doi.org/10.1007/s00170-015-7809-4

    Wu, H., Wang, Y., Yu, Z.: In situ monitoring of fdm machine condition via acous- tic emission. The International Journal of Advanced Manufacturing Technology 84, 1483–1495 (2016) https://doi.org/10.1007/s00170-015-7809-4

    2016 doi

  18. [26]

    Additive Manufacturing21, 598–604 (2018) https://doi.org/10

    Shevchik, S.A., Kenel, C., Leinenbach, C., Wasmer, K.: Acoustic emission for in situ quality monitoring in additive manufacturing using spectral convolutional neural networks. Additive Manufacturing21, 598–604 (2018) https://doi.org/10. 1016/j.addma.2017.11.012

    2018

  19. [27]

    Additive Manufacturing61, 103344 (2023) https://doi.org/10.1016/j.addma.2022.103344

    Pierre, J., Iervolino, F., Farahani, R.D., Piccirelli, N., L´ evesque, M., Therriault, D.: Material extrusion additive manufacturing of multifunctional sandwich panels with load-bearing and acoustic capabilities for aerospace applications. Additive Manufacturing61, 103344 (202...

    2023

  20. [28]

    Progress in Additive Manufacturing4, 211–223 (2019) https://doi.org/10.1007/s40964-019-00089-3

    Tlegenov, Y., Lu, W.F., Hong, G.S.: A dynamic model for current-based noz- zle condition monitoring in fused deposition modelling. Progress in Additive Manufacturing4, 211–223 (2019) https://doi.org/10.1007/s40964-019-00089-3

    2019 doi

  21. [29]

    Additive Manufacturing 56, 102838 (2022) https://doi.org/10.1016/j.addma.2022.102838

    Brion, D.A.J., Shen, M., Pattinson, S.W.: Automated recognition and correction of warp deformation in extrusion additive manufacturing. Additive Manufacturing 56, 102838 (2022) https://doi.org/10.1016/j.addma.2022.102838

    2022

  22. [30]

    Nature Communications13, 4654 (2022) https://doi.org/10.1038/s41467-022-31985-y

    Brion, D.A.J., Pattinson, S.W.: Generalisable 3d printing error detection and cor- rection via multi-head neural networks. Nature Communications13, 4654 (2022) https://doi.org/10.1038/s41467-022-31985-y

    2022 doi

  23. [31]

    Advanced Intelligent Systems4, 2200153 32 (2022) https://doi.org/10.1002/aisy.202200153

    Brion, D.A.J., Pattinson, S.W.: Quantitative and real-time control of 3d printing material flow through deep learning. Advanced Intelligent Systems4, 2200153 32 (2022) https://doi.org/10.1002/aisy.202200153

    2022 doi

  24. [32]

    Additive Manufacturing89, 104271 (2024) https://doi.org/ 10.1016/j.addma.2024.104271

    Margadji, C., Brion, D.A.J., Pattinson, S.W.: Iterative learning for efficient addi- tive mass production. Additive Manufacturing89, 104271 (2024) https://doi.org/ 10.1016/j.addma.2024.104271

    2024

  25. [33]

    https://arxiv

    Margadji, C., Kuswoyo, A., Pattinson, S.W.: Regularized interpolation in 4D neural fields enables optimization of 3D printed geometries (2025). https://arxiv. org/abs/2502.01517

    2025 arXiv

  26. [34]

    https://arxiv.org/abs/2407.21783

    Grattafiori, A., Dubey, A., Jauhri, A., Pandey, A., Kadian, A., Al-Dahle, A., Let- man, A., Mathur, A., Schelten, A., Vaughan, A., Yang, A., Fan, A., Goyal, A., Hartshorn, A., Yang, A., Mitra, A., Sravankumar, A., Korenev, A., Hinsvark, A., Rao, A., Zhang, A., Rodriguez, A., G...

    2024 arXiv

  27. [35]

    2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770–778 (2015)

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 770–778 (2015)

    2015

  28. [36]

    https:// arxiv.org/abs/2106.09685

    Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W.: LoRA: Low-Rank Adaptation of Large Language Models (2021). https:// arxiv.org/abs/2106.09685

    2021 arXiv

  29. [37]

    https://arxiv.org/abs/1606.05250

    Rajpurkar, P., Zhang, J., Lopyrev, K., Liang, P.: SQuAD: 100,000+ Questions for Machine Comprehension of Text (2016). https://arxiv.org/abs/1606.05250

    2016 arXiv

  30. [38]

    https://arxiv.org/abs/1505.04870

    Plummer, B.A., Wang, L., Cervantes, C.M., Caicedo, J.C., Hockenmaier, J., Lazebnik, S.: Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Models (2016). https://arxiv.org/abs/1505.04870

    2016 arXiv

  31. [39]

    Elo, A.E.: The rating of chessplayers, past and present. (1978). https://api.semanticscholar.org/CorpusID:142610973

    1978

  32. [40]

    ArXiv abs/2305.02463(2023)

    Jun, H., Nichol, A.: Shap-e: Generating conditional 3d implicit functions. ArXiv abs/2305.02463(2023)

    2023 arXiv

  33. [41]

    https://arxiv.org/abs/2304.02643 35

    Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A.C., Lo, W.-Y., Doll´ ar, P., Girshick, R.: Segment Anything (2023). https://arxiv.org/abs/2304.02643 35

    2023 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.