pith. sign in

arxiv: 2606.08238 · v1 · pith:RAGMW5JTnew · submitted 2026-06-06 · 💻 cs.LG

GPT-Micro: A large language paradigm for accelerated, inexpensive, and thermodynamics-consistent discovery of constitutive models in manufacturing

Soumik Dutta , Kiarash Naghavi Khanghah , Sania Shree , Logan McNeil , Thomas Feldhausen , Hongyi Xu , Rajiv Malhotra This is my paper

Pith reviewed 2026-06-27 20:05 UTC · model grok-4.3

classification 💻 cs.LG
keywords constitutive modelslarge language modelsthermodynamics compliancemanufacturing processesmodel discoverysparse dataprinted electronicsphysics-informed modeling
0
0 comments X

The pith

GPT-Micro uses large language models to autonomously discover thermodynamics-compliant constitutive models for manufacturing with over 70 percent less data and 400 times faster than prior approaches.

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

The paper presents GPT-Micro as a framework that combines literature-based knowledge extraction, strict enforcement of thermodynamic conservation laws, and sparse experimental datasets with LLM-driven generation and iterative refinement of model functional forms. Traditional constitutive modeling relies on human postulation of forms, which is slow and prone to error, while standard machine learning demands large data volumes and often produces black-box results that ignore physics. By removing the need for a human-chosen starting hypothesis and embedding physics compliance directly into the discovery loop, the approach yields compact analytical models validated on a printed-electronics process. A reader would care because such models directly influence microstructure control in manufacturing, where faster and lower-cost discovery could improve process reliability without sacrificing physical fidelity.

Core claim

GPT-Micro integrates semantic extraction from literature, thermodynamics-based conservation laws, and sparse datasets with LLM generation and refinement of hypotheses to produce de-novo constitutive models. Validation on a printed-electronics testbed shows simultaneous gains: more than 70 percent reduction in data burden relative to ML methods without accuracy loss, 400X reduction in discovery time from months to hours relative to human-driven modeling, novel functional forms without subjective human starting hypotheses, and enhanced trustworthiness through compact, conservation-compliant analytical expressions.

What carries the argument

The GPT-Micro paradigm, which couples LLM hypothesis generation and refinement with automated enforcement of thermodynamic conservation laws on sparse data and literature-derived knowledge.

If this is right

  • More than 70 percent reduction in data burden relative to ML-based modeling without loss in accuracy.
  • 400X reduction in discovery time after data generation, from months to hours, relative to human-driven modeling.
  • Discovery of models with novel functional forms without subjective human choice of a starting hypothesis.
  • Enhanced physics-rooted trustworthiness, human interpretability, and mechanistic insight via synthesis of compact, conservation-compliant, and physically complete analytical models.

Where Pith is reading between the lines

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

  • If the paradigm holds, it could be applied to constitutive modeling in other manufacturing domains such as metal forming or polymer processing where data are similarly expensive.
  • The approach opens the possibility of coupling the LLM loop directly to automated experimental platforms for closed-loop model refinement.
  • Success would imply that literature semantics plus physics constraints can substitute for large labeled datasets in generating interpretable scientific models.

Load-bearing premise

An LLM can reliably generate and refine functional forms that are novel yet strictly compliant with thermodynamic laws, and that metrics on the printed-electronics testbed reflect generalization beyond the sparse training data.

What would settle it

A generated model that violates a thermodynamic conservation law on an independent simulation set, or that shows prediction error increasing sharply when tested on data drawn from a different region of the process parameter space than the sparse training set.

Figures

Figures reproduced from arXiv: 2606.08238 by Hongyi Xu, Kiarash Naghavi Khanghah, Logan McNeil, Rajiv Malhotra, Sania Shree, Soumik Dutta, Thomas Feldhausen.

Figure 1
Figure 1. Figure 1: Components of computational models for microstructure prediction and control. 5-7 Significant effort has been expended on solution methods. Such methods include direct numerical simulations such as Finite Element Analysis (FEA8, 9), Molecular Dynamics10, Phase-Field Modeling11, and Cellular Automata12; and Machine Learning techniques like Physics Informed Neural Networks (PINNs)5 , Deep Operator Neural Net… view at source ↗
Figure 2
Figure 2. Figure 2: Schematic of scalability-data tradeoff in the state-of-the-art constitutive modeling methods. The major approaches for constitutive modeling in the state of the art are as follows: (1) Human-driven modeling: The traditional human-driven approach synthesizes a model form based on intuition, expert interpretation of the literature, and a small amount of high-fidelity data (Fig. 2a). This dependence on human … view at source ↗
Figure 4
Figure 4. Figure 4: Workflow for autonomous knowledge retrieval. [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Workflow for autonomous hypothesis generation and refinement. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: (a) Experimental SEM micrographs of nanowire sintering process [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: (a) R2 score and (b) MSE for state-microstructure models. Note that equations 4 and 5 are merely expressions of the constitutive laws and no human postulation of the form of state-microstructure models or constitutive models is required, unlike conventional human-driven modeling. Further, no human choice of legacy models or datasets is required, unlike Domain Adaptation techniques. Thus, GPT-Micro goes bey… view at source ↗
Figure 8
Figure 8. Figure 8: R2 score evolution with hypothesis refinement Table S2 indicates that the LLM makes some unfounded and vague assumptions regarding the dependence of 𝜃̇ on material states. For example, point 1 of Table S2 states that greater shrinkage (coalescence) due to greater atomic mobility increases the inter-nanowire rotation. This is a hallucination, given the lack of literature on inter-nanowire rotation in the pa… view at source ↗
Figure 9
Figure 9. Figure 9: Comparison of accuracy and data burden for GPT [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
read the original abstract

Constitutive modeling of the relationship between process-imposed material states and fundamental material properties is critical to control of material microstructure in manufacturing processes. The limited accuracy resulting from the typical reliance on fallible human expertise and intuition for postulation and revision of the models functional form results in incremental and time consuming model discovery. Conventional Machine Learning (ML) incurs significant cost and time of data generation. Model discovery using Large Language Models (LLMs) suffers from the above issues and/or ignores the inviolability of fundamental thermodynamics laws. This work creates a novel GPT-Micro paradigm for autonomous, data sparse, and thermodynamics-compliant discovery of de-novo constitutive models. This framework seamlessly integrates semantic knowledge extraction from literature, enforcement of thermodynamics-based conservation laws, and sparse datasets, with LLM-driven generation and refinement of model hypotheses. Validation is performed for a long-intractable constitutive modeling problem in a printed electronics process testbed. This reveals significant and simultaneous advantages over the state-of-the-art including: (a) More than 70 percent reduction in data burden relative to ML-based modeling without loss in accuracy; (b) 400X reduction in discovery time after data generation, from months to hours, relative to human-driven modeling; (c) Discovery of models with novel functional forms without subjective human choice of a starting hypothesis; (d) Enhanced physics-rooted trustworthiness, human interpretability, and mechanistic insight via synthesis of compact, conservation-compliant, and physically complete analytical models. The potential of GPT-Micro to realize rapid, low-cost, physically trustworthy, and interpretable microstructure modeling across the manufacturing landscape is discussed.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript introduces GPT-Micro, an LLM-driven framework that integrates semantic extraction from literature, enforcement of thermodynamic conservation laws, and sparse experimental datasets to autonomously generate and refine de-novo constitutive models. It claims validation on a printed-electronics testbed yielding >70% reduction in data burden, 400X reduction in discovery time, novel functional forms, and enhanced physics compliance relative to human-driven or conventional ML approaches.

Significance. If the thermodynamic enforcement mechanism is shown to be independent and non-circular, the work could meaningfully advance constitutive modeling in manufacturing by lowering data requirements while preserving interpretability and physical consistency. The combination of LLM hypothesis generation with explicit conservation-law constraints addresses a recognized bottleneck, though the absence of detailed enforcement equations and quantitative baselines limits immediate assessment of the claimed gains.

major comments (3)
  1. [Methodology] Methodology (thermodynamic enforcement subsection): the claim of 'strictly compliant' models with thermodynamic laws is load-bearing for the central contribution, yet the text provides no explicit equations, pseudocode, or independent symbolic verifier (e.g., automated check of the Clausius-Duhem inequality) for how compliance is enforced inside the LLM generation/refinement loop; reliance on prompting or self-critique alone risks circularity as noted in the stress-test.
  2. [Results] Results (validation on printed-electronics testbed): the reported 70% data reduction and 400X time reduction lack error bars, statistical tests, or direct quantitative baselines against standard symbolic regression or ML constitutive models on the same sparse dataset, making it impossible to confirm the performance claims are not artifacts of the chosen testbed.
  3. [Abstract and §4] Abstract and §4 (model discovery claims): the assertion of 'novel functional forms without subjective human choice' is not supported by an ablation showing that the LLM-generated forms differ systematically from those reachable by conventional human postulation or existing symbolic methods on the identical data.
minor comments (2)
  1. Notation for thermodynamic quantities is introduced without a consolidated table of symbols, complicating cross-referencing between the enforcement description and the discovered model equations.
  2. Figure captions for the testbed results do not state the number of independent runs or the precise definition of 'discovery time' (wall-clock after data generation vs. total including prompting iterations).

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which identify key areas for strengthening the manuscript. We address each major comment below and will incorporate revisions to provide the requested details, baselines, and ablations.

read point-by-point responses

  1. Referee: [Methodology] Methodology (thermodynamic enforcement subsection): the claim of 'strictly compliant' models with thermodynamic laws is load-bearing for the central contribution, yet the text provides no explicit equations, pseudocode, or independent symbolic verifier (e.g., automated check of the Clausius-Duhem inequality) for how compliance is enforced inside the LLM generation/refinement loop; reliance on prompting or self-critique alone risks circularity as noted in the stress-test.

    Authors: We agree that explicit documentation of the enforcement mechanism is required to substantiate the central claim of strict thermodynamic compliance. The revised manuscript will add a dedicated subsection containing the explicit mathematical formulation of the Clausius-Duhem inequality enforcement, pseudocode for its integration into the LLM generation and refinement loop, and a description of the independent symbolic verifier. These additions will demonstrate that compliance checking occurs via an external, non-circular procedure rather than relying solely on LLM prompting or self-critique. revision: yes

  2. Referee: [Results] Results (validation on printed-electronics testbed): the reported 70% data reduction and 400X time reduction lack error bars, statistical tests, or direct quantitative baselines against standard symbolic regression or ML constitutive models on the same sparse dataset, making it impossible to confirm the performance claims are not artifacts of the chosen testbed.

    Authors: We accept that the performance metrics require statistical rigor and direct baselines for credibility. The revision will include error bars on all reported metrics, statistical significance tests (e.g., paired t-tests), and quantitative comparisons against symbolic regression (such as PySR) and conventional ML models evaluated on the identical sparse printed-electronics dataset. These results will be presented in an expanded results section and accompanying table. revision: yes

  3. Referee: [Abstract and §4] Abstract and §4 (model discovery claims): the assertion of 'novel functional forms without subjective human choice' is not supported by an ablation showing that the LLM-generated forms differ systematically from those reachable by conventional human postulation or existing symbolic methods on the identical data.

    Authors: This observation correctly identifies the need for empirical support of the novelty claim. We will revise §4 to include an ablation study that applies human postulation and existing symbolic regression methods to the same dataset and compares the resulting functional forms with those generated by GPT-Micro. The abstract will be updated to reference this evidence of systematic differences. revision: yes

Circularity Check

0 steps flagged

No circularity: framework claims rest on external integration rather than self-referential definitions or fits

full rationale

The abstract and description present GPT-Micro as integrating literature extraction, thermodynamics enforcement, sparse data, and LLM hypothesis generation/refinement, with validation on a printed-electronics testbed. No equations, parameter-fitting steps, or self-citations are provided that would reduce any performance claim (e.g., 70% data reduction or thermodynamics compliance) to a tautology by construction. The thermodynamics enforcement is described as a distinct module rather than a post-hoc filter or loss tuned on the same outputs, and no uniqueness theorems or ansatzes from prior author work are invoked to force the result. The derivation chain therefore remains self-contained against the stated inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies no explicit free parameters, axioms, or invented entities; the framework itself (GPT-Micro) is the central new construct whose internal mechanics are not described.

pith-pipeline@v0.9.1-grok · 5855 in / 1286 out tokens · 16776 ms · 2026-06-27T20:05:13.567437+00:00 · methodology

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

62 extracted references · 1 linked inside Pith

  1. [1]

    Journal of Materials Research and Technology 2022, 18, 3056-3072

    Tao Zhang; Huigui Li; Hai Gong; Yunxin Wu; Xin Chen; Xiaoyong Zhang, Study on location- related thermal cycles and microstructure variation of additively manufactured inconel 718. Journal of Materials Research and Technology 2022, 18, 3056-3072

    2022

  2. [2]

    Additive Manufacturing 2021, 46, 102033

    Francis Ogoke; Amir Barati Farimani, Thermal control of laser powder bed fusion using deep reinforcement learning. Additive Manufacturing 2021, 46, 102033

    2021

  3. [3]

    The Influence of Cooling Rate on Microstructure and Mechanical Properties of AlSi9Cu3 Metals [Online], 2021

    Matic Žbontar; Mitja Petrič; Primož Mrvar. The Influence of Cooling Rate on Microstructure and Mechanical Properties of AlSi9Cu3 Metals [Online], 2021

    2021

  4. [4]

    Procedia Structural Integrity 2023, 43, 240-245

    Matthias Oberreiterr; Michael Stoschka; Michael Horvath; Stefan Fladischer, Effect of local microstructure on fatigue and related failure mechanisms in AlSi-castings. Procedia Structural Integrity 2023, 43, 240-245

    2023

  5. [5]

    Computational Mechanics 2021, 67, 619-635

    Qiming Zhu; Zeliang Liu; Jinhui Yan, Machine learning for metal additive manufacturing: predicting temperature and melt pool fluid dynamics using physics-informed neural networks. Computational Mechanics 2021, 67, 619-635

    2021

  6. [6]

    Nature Machine Intelligence 2021, 3 (3), 218-229

    Lu Lu; Pengzhan Jin; Guofei Pang; Zhongqiang Zhang; George Em Karniadakis, Learning nonlinear operators via DeepONet based on the universal approximation theorem of operators. Nature Machine Intelligence 2021, 3 (3), 218-229

    2021

  7. [7]

    npj Computational Materials 2022, 8 (1), 201

    Tianju Xue; Zhengtao Gan; Shuheng Liao; Jian Cao, Physics-embedded graph network for accelerating phase-field simulation of microstructure evolution in additive manufacturing. npj Computational Materials 2022, 8 (1), 201

    2022

  8. [8]

    Journal of Materials Processing Technology 2026, 352, 119316

    Tobias Cyra-Wolf; Amir Malakizadi; Jannis Saelzer; Youssef Alammari; Dirk Biermann, A physics- based flow stress model for chip formation simulation when machining Alloy 718. Journal of Materials Processing Technology 2026, 352, 119316

    2026

  9. [9]

    Journal of Materials Processing Technology 2026, 348, 119185

    Arvind Chouhan; Lutz Mädler; Nils Ellendt, Modeling composition-dependent melt dynamics and defect formation in multi-material additive manufacturing. Journal of Materials Processing Technology 2026, 348, 119185

    2026

  10. [10]

    Journal of Materials Processing Technology 2026, 347, 119145

    Zhicheng Chen; Jingya Sun; Gen Lin; Ziqi Ding; Yang Ding; Lan Jiang, Atomistic insights into ultrafast phase transitions driving nanoparticle formation in femtosecond laser–irradiated nickel-based superalloys. Journal of Materials Processing Technology 2026, 347, 119145

    2026

  11. [11]

    Journal of Materials Research and Technology 2024, 28, 4585-4599

    Fuchen Wang; Weipeng Chen; Dong Wang; Hua Hou; Yuhong Zhao, Phase-field modeling and Experimental investigation for rapid solidification in wire and arc additive manufacturing. Journal of Materials Research and Technology 2024, 28, 4585-4599

    2024

  12. [12]

    Rowenhorst, Optimizing the cellular automata finite element model for additive manufacturing to simulate large microstructures

    Kirubel Teferra; David J. Rowenhorst, Optimizing the cellular automata finite element model for additive manufacturing to simulate large microstructures. Acta Materialia 2021, 213, 116930

    2021

  13. [13]

    M. W. Reiterer; K. G. Ewsuk; J. G. Argüello, An Arrhenius-Type Viscosity Function to Model Sintering Using the Skorohod–Olevsky Viscous Sintering Model Within a Finite-Element Code. Journal of the American Ceramic Society 2006, 89 (6), 1930-1935

    2006

  14. [14]

    Scripta Materialia 2006, 55 (12), 1175-1178

    Eugene Olevsky; Ludo Froyen, Constitutive modeling of spark-plasma sintering of conductive materials. Scripta Materialia 2006, 55 (12), 1175-1178

    2006

  15. [15]

    Kooi; Yutao Pei, Additive manufactured high entropy alloys: A review of the microstructure and properties

    Wei Zhang; Ali Chabok; Bart J. Kooi; Yutao Pei, Additive manufactured high entropy alloys: A review of the microstructure and properties. Materials & Design 2022, 220, 110875. 21

    2022

  16. [16]

    Sarah J. Wolff; Benjamin Gould; Niranjan Parab; Cang Zhao; Aaron Greco; Tao Sun, Preliminary Study on the Influence of an External Magnetic Field on Melt Pool Behavior in Laser Melting of 4140 Steel Using In-Situ X-Ray Imaging. Journal of Micro and Nano-Manufacturing 2021, 8 (4)

    2021

  17. [17]

    Journal of Manufacturing Processes 2024, 127, 328-339

    Jianwu Huang; Haiou Zhang; Runsheng Li; Xushan Zhao; Hang Lin; Wenzheng Zhai; Guilan Wang; Youheng Fu, Hybrid in-situ hot rolling and wire arc additive manufacturing of Al-Si alloy: Microstructure, mechanical properties and strengthening mechanism. Journal of Manufacturing Processes 2024, 127, 328-339

    2024

  18. [18]

    Journal of Manufacturing Processes 2023, 99, 848-897

    Katayoon Taherkhani; Osazee Ero; Farima Liravi; Sahar Toorandaz; Ehsan Toyserkani, On the application of in-situ monitoring systems and machine learning algorithms for developing quality assurance platforms in laser powder bed fusion: A review. Journal of Manufacturing Processes 2023, 99, 848-897

    2023

  19. [19]

    Joost R. Duflou; Anne-Marie Habraken; Jian Cao; Rajiv Malhotra; Markus Bambach; Dave Adams; Hans Vanhove; Amirahmad Mohammadi; Jack Jeswiet, Single point incremental forming: state- of-the-art and prospects. International Journal of Material Forming 2018, 11 (6), 743-773

    2018

  20. [20]

    Ehmann; Jian Cao, Prediction of forming temperature in electrically-assisted double-sided incremental forming using a neural network

    Zilin Jiang; Kornel F. Ehmann; Jian Cao, Prediction of forming temperature in electrically-assisted double-sided incremental forming using a neural network. Journal of Materials Processing Technology 2022, 302, 117486

    2022

  21. [21]

    Journal of Materials Research and Technology 2023, 22, 2720-2735

    Yiwei Mao; Chao Cai; Jinkai Zhang; Yuhua Heng; Kunhao Feng; Daosheng Cai; Qingsong Wei, Effect of sintering temperature on binder jetting additively manufactured stainless steel 316L: densification, microstructure evolution and mechanical properties. Journal of Materials Research and Technology 2023, 22, 2720-2735

    2023

  22. [22]

    Łukasz Łach. Cellular Automata and Phase-Field Modeling of Microstructure Evolution in Metal Additive Manufacturing: Recent Advances, Hybrid Frameworks, and Pathways to Predictive Control Metals [Online], 2026, p. 124

    2026

  23. [23]

    Journal of Materials Research and Technology 2025, 36, 3974-3987

    Jiale Wang; Bo Yin; Kai Qin; Haohao Miao; Yu Yin; Petro Pavlenko; Yuefeng Yuan; Jinbao Wang; David Henry StJohn; Zhao Zhang; Hao Yi, A new analytical model for process-microstructure- hardness relationships of directed energy deposited nickel alloys. Journal of Materials Research and Technology 2025, 36, 3974-3987

    2025

  24. [24]

    Joshi; K

    Shashank Sharma; Saikumar Dussa; Sameehan S. Joshi; K. V. Mani Krishna; Raj Banerjee; Narendra B. Dahotre, Thermokinetics-driven evolution of grain morphologies during laser direct energy deposition. Journal of Manufacturing Processes 2025, 151, 895-915

    2025

  25. [25]

    Additive Manufacturing 2022, 58, 103047

    Tao Zhou; Tao Zheng; Ahmet Bahadir Yildiz; Gabriel Spartacus; Monika Rolinska; Robert Cubitt; Peter Hedström, Microstructure control during deposition and post-treatment to optimize mechanical properties of wire-arc additively manufactured 17-4 PH stainless steel. Additive Manufacturing 2022, 58, 103047

    2022

  26. [26]

    Journal of Materials Informatics 2023, 3 (1), 7

    Menon Nandana; Mondal Sudeepta; Basak Amrita, Linking processing parameters with melt pool properties of multiple nickel-based superalloys via high-dimensional Gaussian process regression. Journal of Materials Informatics 2023, 3 (1), 7

    2023

  27. [27]

    Journal of Manufacturing Processes 2022, 79, 803-814

    Noopur Dilip Jamnikar; Sen Liu; Craig Brice; Xiaoli Zhang, In situ microstructure property prediction by modeling molten pool-quality relations for wire-feed laser additive manufacturing. Journal of Manufacturing Processes 2022, 79, 803-814

    2022

  28. [28]

    Virtual and Physical Prototyping 2022, 17 (2), 382-405

    Ze Chen; Changjun Han; Ming Gao; Sastry Yagnanna Kandukuri; Kun Zhou, A review on qualification and certification for metal additive manufacturing. Virtual and Physical Prototyping 2022, 17 (2), 382-405

    2022

  29. [29]

    Journal of Manufacturing Science and Engineering 2025, 147 (5)

    Manaswin Oddiraju; Jeremy Cleeman; Rajiv Malhotra; Souma Chowdhury, A Differentiable Physics-Informed Machine Learning Approach to Model Laser-Based Micro-Manufacturing Process. Journal of Manufacturing Science and Engineering 2025, 147 (5). 22

    2025

  30. [30]

    Gary Wang, Selecting subsets of source data for transfer learning with applications in metal additive manufacturing

    Yifan Tang; Mostafa Rahmani Dehaghani; Pouyan Sajadi; G. Gary Wang, Selecting subsets of source data for transfer learning with applications in metal additive manufacturing. Journal of Intelligent Manufacturing 2024, 1-22

    2024

  31. [31]

    Journal of Materials Processing Technology 2022, 303, 117531

    Vigneashwara Pandiyan; Rita Drissi-Daoudi; Sergey Shevchik; Giulio Masinelli; Tri Le-Quang; Roland Loge; Kilian Wasmer, Deep transfer learning of additive manufacturing mechanisms across materials in metal-based laser powder bed fusion process. Journal of Materials Processing Technology 2022, 303, 117531

    2022

  32. [32]

    Journal of Materials Processing Technology 2023, 320, 118125

    Jeremy Cleeman; Kian Agrawala; Evan Nastarowicz; Rajiv Malhotra, Partial-physics-informed multi-fidelity modeling of manufacturing processes. Journal of Materials Processing Technology 2023, 320, 118125

    2023

  33. [33]

    Advanced Engineering Informatics 2026, 74, 104707

    Anandkumar Patel; Jeremy Cleeman; Rajiv Malhotra, Source tolerant iterative domain adaptation for inexpensive modeling of knowledge-deficient manufacturing processes. Advanced Engineering Informatics 2026, 74, 104707

    2026

  34. [34]

    Milani, Making costly manufacturing smart with transfer learning under limited data: A case study on composites autoclave processing

    Milad Ramezankhani; Bryn Crawford; Apurva Narayan; Heinz Voggenreiter; Rudolf Seethaler; Abbas S. Milani, Making costly manufacturing smart with transfer learning under limited data: A case study on composites autoclave processing. Journal of Manufacturing Systems 2021, 59, 345-354

    2021

  35. [35]

    Saunders; Kirubel Teferra; Alaa Elwany; John G

    Robert N. Saunders; Kirubel Teferra; Alaa Elwany; John G. Michopoulos; Dimitris Lagoudas, Metal AM process-structure-property relational linkages using Gaussian process surrogates. Additive Manufacturing 2023, 62, 103398

    2023

  36. [36]

    Integrating Materials and Manufacturing Innovation 2022, 1- 19

    Robert Saunders; Anna Rawlings; Andrew Birnbaum; Athanasios Iliopoulos; John Michopoulos; Dimitris Lagoudas; Alaa Elwany, Additive manufacturing melt pool prediction and classification via multifidelity gaussian process surrogates. Integrating Materials and Manufacturing Innovation 2022, 1- 19

    2022

  37. [37]

    Precision Engineering 2020, 62, 196-203

    Takashi Misaka; Jonny Herwan; Oleg Ryabov; Seisuke Kano; Hiroyuki Sawada; Nagayoshi Kasashima; Yoshiyuki Furukawa, Prediction of surface roughness in CNC turning by model-assisted response surface method. Precision Engineering 2020, 62, 196-203

    2020

  38. [38]

    I. T. Tandogan; M. Budnitzki; S. Sandfeld, A multi-physics model for the evolution of grain microstructure. International Journal of Plasticity 2025, 185, 104201

    2025

  39. [39]

    Genetic Programming and Evolvable Machines 2024, 26 (1), 7

    Alberto Tonda, Review of PySR: high-performance symbolic regression in Python and Julia. Genetic Programming and Evolvable Machines 2024, 26 (1), 7

    2024

  40. [40]

    Journal of Manufacturing Processes 2020, 58, 1088-1102

    Harish Devaraj; Hyun-Jun Hwang; Rajiv Malhotra, Understanding the role of Nanomorphology on Resistance Evolution in the Hybrid Form-Fuse Process for Conformal Electronics. Journal of Manufacturing Processes 2020, 58, 1088-1102

    2020

  41. [41]

    Allen; Mikko Aronniemi; Tomi Mattila; Ari Alastalo; Kimmo Ojanperä; Mika Suhonen; Heikki Seppä, Electrical sintering of nanoparticle structures

    Mark L. Allen; Mikko Aronniemi; Tomi Mattila; Ari Alastalo; Kimmo Ojanperä; Mika Suhonen; Heikki Seppä, Electrical sintering of nanoparticle structures. Nanotechnology 2008, 19 (17), 175201

    2008

  42. [42]

    Rinker; Changbai Tan; Blair Carlson; Siguang Xu; Jingjing Li, Derivation of physical equations for high-speed laser welding using large language models

    Kyubok Lee; Zhengxiao Yu; Zen-Hao Lai; Peihao Geng; Teresa J. Rinker; Changbai Tan; Blair Carlson; Siguang Xu; Jingjing Li, Derivation of physical equations for high-speed laser welding using large language models. International Journal of Machine Tools and Manufacture 2025, 211, 104320

    2025

  43. [43]

    Journal of Intelligent Manufacturing 2026, 37 (5), 2085-2113

    Kiarash Naghavi Khanghah; AnandKumar Patel; Rajiv Malhotra; Hongyi Xu, Large language models for extrapolative modeling of manufacturing processes. Journal of Intelligent Manufacturing 2026, 37 (5), 2085-2113

    2026

  44. [44]

    Buehler, MechGPT, a Language-Based Strategy for Mechanics and Materials Modeling That Connects Knowledge Across Scales, Disciplines, and Modalities

    Markus J. Buehler, MechGPT, a Language-Based Strategy for Mechanics and Materials Modeling That Connects Knowledge Across Scales, Disciplines, and Modalities. Applied Mechanics Reviews 2024, 76 (2)

    2024

  45. [45]

    Machine Learning for Computational Science and Engineering 2025, 1 (1), 23

    Marius Tacke; Matthias Busch; Kartik Bali; Kian Abdolazizi; Kevin Linka; Christian Cyron; Roland Aydin, Constitutive scientific generative agent (CSGA): Leveraging large language models for automated constitutive model discovery. Machine Learning for Computational Science and Engineering 2025, 1 (1), 23. 23

    2025

  46. [46]

    arXiv preprint arXiv:2512.01735 2025

    Marius Tacke; Matthias Busch; Kian Abdolazizi; Jonas Eichinger; Kevin Linka; Christian Cyron; Roland Aydin, Automating modeling in mechanics: LLMs as designers of physics-constrained neural networks for constitutive modeling of materials. arXiv preprint arXiv:2512.01735 2025

    arXiv 2025

  47. [47]

    Computers, Materials and Continua 2026, 87 (2)

    Yuchuan Gu; Lusheng Wang; Jun Ding; Yanhong Peng; Changgeng Li; Shaojie Gu, Large Language Model-Enabled Constitutive Modeling for Rate-Dependent Plasticity and Automatic UMAT Subroutine Generation. Computers, Materials and Continua 2026, 87 (2)

    2026

  48. [48]

    Rinker; Blair Carlson; Masoud M

    Zhengxiao Yu; Peihao Geng; Bo Pan; Teresa J. Rinker; Blair Carlson; Masoud M. Pour; Hui Wang; Nan Li; Shun-Li Shang; Zi-Kui Liu; Jingjing Li, LLM-driven analysis of micro-pillar experiments for investigating fracture in Al-steel RSW joints. Journal of Manufacturing Processes 2025, 155, 868-879

    2025

  49. [49]

    Additive Manufacturing Letters 2024, 11, 100232

    Achuth Chandrasekhar; Jonathan Chan; Francis Ogoke; Olabode Ajenifujah; Amir Barati Farimani, AMGPT: A large language model for contextual querying in additive manufacturing. Additive Manufacturing Letters 2024, 11, 100232

    2024

  50. [50]

    arXiv preprint arXiv:2406.05410 2024

    Yanjie Li; Lina Yu; Weijun Li; Min Wu; Liping Zhang; Jingyi Liu; Yusong Deng; Mingzhu Wan; Xin Ning, ChatSR: Multimodal Large Language Models for Scientific Formula Discovery. arXiv preprint arXiv:2406.05410 2024

    Pith/arXiv arXiv 2024

  51. [51]

    Physics of Fluids 2024, 36 (9), 097121

    Mengge Du; Yuntian Chen; Zhongzheng Wang; Longfeng Nie; Dongxiao Zhang, Large language models for automatic equation discovery of nonlinear dynamics. Physics of Fluids 2024, 36 (9), 097121

    2024

  52. [52]

    Advances in Neural Information Processing Systems 2024, 37, 44678-44709

    Arya Grayeli; Atharva Sehgal; Omar Costilla-Reyes; Miles Cranmer; Swarat Chaudhuri, Symbolic regression with a learned concept library. Advances in Neural Information Processing Systems 2024, 37, 44678-44709

    2024

  53. [53]

    In Artificial Intelligence Applications and Innovations, 2024; pp 74-87

    Nicholas Ampazis, Improving RAG Quality for Large Language Models with Topic-Enhanced Reranking. In Artificial Intelligence Applications and Innovations, 2024; pp 74-87

    2024

  54. [54]

    Small" Language Models to Say

    Xinxi Chen; Li Wang; Wei Wu; Qi Tang; Yiyao Liu, Honest AI: Fine-Tuning "Small" Language Models to Say "I Don't Know", and Reducing Hallucination in RAG. 2024, (13 October 2024)

    2024

  55. [55]

    Bansal; R

    S. Bansal; R. Malhotra, Nanoscale-shape-mediated coupling between temperature and densification in intense pulsed light sintering. Nanotechnology 2016, 27 (49), 495602

    2016

  56. [56]

    ACS Applied Materials & Interfaces 2019, 11 (3), 3536-3546

    Hyun-Jun Hwang; Rajiv Malhotra, Shape-Tuned Junction Resistivity and Self-Damping Dynamics in Intense Pulsed Light Sintering of Silver Nanostructure Films. ACS Applied Materials & Interfaces 2019, 11 (3), 3536-3546

    2019

  57. [57]

    CIRP Annals 2026

    Rajiv Malhotra; Md Naim Jahangir; Hongyi Xu, A Print-Form-Sinter process for manufacturing 3D electronics on complex surfaces. CIRP Annals 2026

    2026

  58. [58]

    Additive Manufacturing 2019, 30, 100886

    Md Naim Jahangir; Jeremy Cleeman; Hyun-Jun Hwang; Rajiv Malhotra, Towards out-of-chamber damage-free fabrication of highly conductive nanoparticle-based circuits inside 3D printed thermally sensitive polymers. Additive Manufacturing 2019, 30, 100886

    2019

  59. [59]

    The Journal of Physical Chemistry C 2020, 124 (36), 19849-19857

    Md Naim Jahangir; Harish Devaraj; Rajiv Malhotra, On Self-Limiting Rotation and Diffusion Mechanisms during Sintering of Silver Nanowires. The Journal of Physical Chemistry C 2020, 124 (36), 19849-19857

    2020

  60. [60]

    Advanced Theory and Simulations 2021, 4 (8), 2100104

    Harish Devaraj; Md Naim Jahangir; Zhongwei Gao; Chih-hung Chang; Rajiv Malhotra, Fusion of Stacked Nanowires: From Atomistic to Analytical Models. Advanced Theory and Simulations 2021, 4 (8), 2100104

    2021

  61. [61]

    ACS Applied Materials & Interfaces 2021, 13 (47), 56645- 56654

    Harish Devaraj; Qi Tian; Weihong Guo; Rajiv Malhotra, Multiscale Modeling of Sintering-Driven Conductivity in Large Nanowire Ensembles. ACS Applied Materials & Interfaces 2021, 13 (47), 56645- 56654

    2021

  62. [62]

    Ho; Toshitake Takahashi; Roie Yerushalmi; Kuniharu Takei; Alexandra C

    Zhiyong Fan; Johnny C. Ho; Toshitake Takahashi; Roie Yerushalmi; Kuniharu Takei; Alexandra C. Ford; Yu-Lun Chueh; Ali Javey, Toward the development of printable nanowire electronics and sensors. Advanced Materials 2009, 21 (37), 3730-3743

    2009