pith. sign in

arxiv: 2508.07798 · v1 · submitted 2025-08-11 · ❄️ cond-mat.mtrl-sci · cs.LG

Generative Inversion for Property-Targeted Materials Design: Application to Shape Memory Alloys

Cheng Li , Pengfei Danga , Yuehui Xiana , Yumei Zhou , Bofeng Shi , Xiangdong Ding , Jun Suna , Dezhen Xue This is my paper

Pith reviewed 2026-05-19 00:06 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci cs.LG
keywords shape memory alloysGAN inversionmaterials designinverse designNiTi alloystransformation temperaturegenerative models
0
0 comments X

The pith

GAN inversion generates NiTi alloy compositions that hit 404°C transformation temperature and 9.9 J/cm³ work output.

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

The paper establishes a framework that couples a pretrained generative adversarial network with a property prediction model to perform inverse design of shape memory alloys. Gradient-based optimization in the latent space directly suggests alloy compositions and processing conditions to hit user-specified targets for transformation temperature and work output. The method is validated by synthesizing and testing five NiTi-based alloys, including one that reaches 404°C with 9.9 J/cm³ work output. A sympathetic reader would care because traditional trial-and-error design of these functional materials is slow and costly for high-temperature applications, while this offers a faster route through data-driven search in composition space.

Core claim

By inverting a pretrained GAN through gradient descent guided by a coupled property model, the authors generate and experimentally realize NiTi-based shape memory alloys with targeted high transformation temperatures and large mechanical work outputs. The standout Ni49.8Ti26.4Hf18.6Zr5.2 alloy exhibits a 404°C transformation temperature, 9.9 J/cm³ work output, 43 J/g enthalpy, and 29°C hysteresis, explained by pronounced transformation volume change and finely dispersed Ti2Ni-type precipitates resulting from sluggish Zr and Hf diffusion.

What carries the argument

Generative adversarial network inversion, which optimizes alloy compositions and parameters by navigating the latent space of a pretrained generator using gradients from a property prediction model.

If this is right

  • The framework allows direct targeting of multiple properties including transformation temperature, mechanical work output, and thermal hysteresis.
  • Experimental validation confirms that the generated alloys can outperform conventional NiTi alloys in transformation temperature.
  • Performance improvements trace to microstructural features such as fine precipitates and large volume changes enabled by slow diffusion of Zr and Hf.
  • The generative approach provides an efficient alternative to trial-and-error synthesis for exploring complex alloy spaces.

Where Pith is reading between the lines

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

  • If the latent space optimization reliably transfers to physical synthesis, the same inversion technique could be applied to inverse design of other functional materials with coupled property targets.
  • Adding physics-informed constraints during optimization might further improve the success rate of synthesized candidates.
  • Training the GAN on expanded datasets could uncover compositions with even higher transformation temperatures or work outputs.

Load-bearing premise

The pretrained GAN and property prediction model are accurate enough that gradient optimization in latent space yields compositions whose synthesized properties match the targets without large errors from unmodeled effects.

What would settle it

Synthesizing the Ni49.8Ti26.4Hf18.6Zr5.2 alloy and measuring its actual transformation temperature and mechanical work output to check whether they reach the targeted 404°C and 9.9 J/cm³.

Figures

Figures reproduced from arXiv: 2508.07798 by Bofeng Shi, Cheng Li, Dezhen Xue, Jun Suna, Pengfei Danga, Xiangdong Ding, Yuehui Xiana, Yumei Zhou.

Figure 1
Figure 1. Figure 1: Schematic of the GAN inversion framework for property-targeted inverse design of shape memory alloys. The framework integrates a pretrained GAN with a supervised property predictor to enable inverse mapping from target performance to alloy design. For inverse design, a random latent vector z is sampled from a 10-dimensional latent space and decoded by the generator to produce candidate design vectors (ci ,… view at source ↗
Figure 2
Figure 2. Figure 2: Evaluation of the generative performance of the trained GAN model. (a) t-SNE visualization of real (orange) and GAN-generated (blue) samples in compositional and processing parameters space, showing strong distributional overlap and indicating that the generator captures the underlying data manifold. (b) Diversity analysis of generated samples as a function of generation size. “Unique” and “Unique & New” s… view at source ↗
Figure 3
Figure 3. Figure 3 [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Property-targeted conditional generation: precise, novel, and diverse designs produced via GAN inversion. (a–b). Violin plots of generated samples targeting (a) martensite start temperature (Ms = 200, 400, 500 ◦C) and (b) mechanical work output (10, 20, 25 J/cm3 ). The narrowing of distributions and alignment of medians with target values demonstrate precise property control, in contrast to the broad basel… view at source ↗
Figure 5
Figure 5. Figure 5: Experimental validation of GAN inversion-designed high-temperature shape memory alloys. (a) Distribution of loss, defined as the mean squared error between predicted and target property values, for randomly generated seed samples versus GAN-inversion-optimized candidates. Inset shows conceptual depiction of the inversion objective in the Ms–workout space . (b) t-SNE projection of alloys designed by GAN inv… view at source ↗
Figure 6
Figure 6. Figure 6: Crystallographic and microstructural origin of high performance in designed Alloy A1. (a) XRD patterns collected at 300 ◦C and 500 ◦C, indicate the presence of B19′ martensite phase at 300 ◦C and B2 Austenite phase at 500 ◦C. (b) BSE image of the alloy, showing a homogenous dispersion of fine precipitates across the matrix. Inset shows a higher-magnification view. (c) Histogram of precipitate size distribu… view at source ↗
read the original abstract

The design of shape memory alloys (SMAs) with high transformation temperatures and large mechanical work output remains a longstanding challenge in functional materials engineering. Here, we introduce a data-driven framework based on generative adversarial network (GAN) inversion for the inverse design of high-performance SMAs. By coupling a pretrained GAN with a property prediction model, we perform gradient-based latent space optimization to directly generate candidate alloy compositions and processing parameters that satisfy user-defined property targets. The framework is experimentally validated through the synthesis and characterization of five NiTi-based SMAs. Among them, the Ni$_{49.8}$Ti$_{26.4}$Hf$_{18.6}$Zr$_{5.2}$ alloy achieves a high transformation temperature of 404 $^\circ$C, a large mechanical work output of 9.9 J/cm$^3$, a transformation enthalpy of 43 J/g , and a thermal hysteresis of 29 {\deg}C, outperforming existing NiTi alloys. The enhanced performance is attributed to a pronounced transformation volume change and a finely dispersed of Ti$_2$Ni-type precipitates, enabled by sluggish Zr and Hf diffusion, and semi-coherent interfaces with localized strain fields. This study demonstrates that GAN inversion offers an efficient and generalizable route for the property-targeted discovery of complex alloys.

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

2 major / 3 minor

Summary. The paper introduces a GAN inversion framework that couples a pretrained generative adversarial network with a property prediction model to enable gradient-based optimization in latent space for inverse design of shape memory alloys targeting high transformation temperatures and mechanical work output. Candidate compositions and processing parameters are generated and experimentally validated through synthesis and characterization of five NiTi-based alloys. The Ni49.8Ti26.4Hf18.6Zr5.2 composition is reported to achieve a transformation temperature of 404°C, mechanical work output of 9.9 J/cm³, transformation enthalpy of 43 J/g, and thermal hysteresis of 29°C, with performance gains attributed to large transformation volume change and finely dispersed Ti2Ni-type precipitates enabled by sluggish diffusion of Zr and Hf.

Significance. If the central claims hold, the work demonstrates a practical route for property-targeted discovery in complex functional alloys by combining generative models with experimental feedback, potentially reducing the trial-and-error burden in SMA design. The synthesis and detailed characterization of five alloys, including one with standout metrics relative to conventional NiTi systems, provides concrete grounding and highlights microstructural mechanisms (volume change and precipitate interfaces) that could inform broader alloy engineering. The approach is generalizable in principle to other multi-component systems where direct forward modeling is challenging.

major comments (2)
  1. [§3 and §4.3] §3 (Framework and Optimization) and §4.3 (Experimental Validation): the central claim that gradient-based latent-space optimization produces compositions satisfying the targets rests on the unverified accuracy of the coupled property predictor outside the training distribution. No quantitative metrics are provided for prediction error, uncertainty, or bias on held-out compositions near the Hf/Zr-rich optimized point (e.g., MAE on transformation temperature or work output), nor is the latent-space distance to the training manifold reported. Without these checks, it remains unclear whether the reported performance of Ni49.8Ti26.4Hf18.6Zr5.2 reflects reliable inversion or post-hoc experimental success.
  2. [Table 2] Table 2 or equivalent results table (alloy performance summary): while the five synthesized alloys are presented with specific numbers, the manuscript does not include direct comparison of model-predicted versus measured properties for the optimized compositions, nor error bars or replicate measurements that would allow assessment of whether the optimization step systematically reduced the property gap relative to random or expert-guided selection.
minor comments (3)
  1. [Abstract] Abstract and §5 (Microstructure discussion): the phrase 'finely dispersed of Ti2Ni-type precipitates' contains a grammatical error; revise for clarity.
  2. [Figure 2] Figure 2 (schematic of GAN inversion pipeline): the diagram would benefit from explicit annotation of the gradient flow during latent-space optimization and the interface between the GAN generator and the property predictor.
  3. [Methods] Methods section: ensure all training hyperparameters for the GAN and property model, as well as the exact composition of the training dataset, are fully specified to support reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive feedback and positive evaluation of the work's significance. We address the major comments below regarding validation of the property predictor and direct comparisons in the results. Revisions will be made to incorporate additional metrics and comparisons where feasible.

read point-by-point responses

  1. Referee: [§3 and §4.3] §3 (Framework and Optimization) and §4.3 (Experimental Validation): the central claim that gradient-based latent-space optimization produces compositions satisfying the targets rests on the unverified accuracy of the coupled property predictor outside the training distribution. No quantitative metrics are provided for prediction error, uncertainty, or bias on held-out compositions near the Hf/Zr-rich optimized point (e.g., MAE on transformation temperature or work output), nor is the latent-space distance to the training manifold reported. Without these checks, it remains unclear whether the reported performance of Ni49.8Ti26.4Hf18.6Zr5.2 reflects reliable inversion or post-hoc experimental success.

    Authors: We agree that explicit quantification of the property predictor's performance outside the training distribution would strengthen confidence in the latent-space optimization. The manuscript prioritizes experimental validation of the generated candidates as the ultimate test of the framework. In the revised manuscript, we will add MAE, bias, and uncertainty estimates for transformation temperature and work output on a held-out test set, including subsets of compositions proximate to the Hf/Zr-rich region. We will also compute and report the latent-space distances of the optimized points relative to the training manifold to characterize the degree of extrapolation. These additions will be placed in §3 or as a new supplementary table. revision: yes

  2. Referee: [Table 2] Table 2 or equivalent results table (alloy performance summary): while the five synthesized alloys are presented with specific numbers, the manuscript does not include direct comparison of model-predicted versus measured properties for the optimized compositions, nor error bars or replicate measurements that would allow assessment of whether the optimization step systematically reduced the property gap relative to random or expert-guided selection.

    Authors: We acknowledge the value of showing model predictions alongside measured values to evaluate the optimization step. The revised manuscript will include a direct comparison table (or expanded Table 2) listing both the model-predicted and experimentally measured properties for all five synthesized alloys. This will allow assessment of prediction fidelity for the optimized candidates. Regarding error bars and replicates, the presented data derive from single synthesis and characterization runs per composition, as is common for resource-intensive alloy development; we will note this limitation explicitly and provide context on typical measurement variability from our prior work on similar systems. We will also add a brief comparison to randomly sampled latent-space points to illustrate the targeted improvement. revision: partial

Circularity Check

0 steps flagged

Pretrained GAN + predictor framework grounded by experimental synthesis; no derivation reduces to fitted inputs by construction

full rationale

The paper's core workflow couples a pretrained GAN with a separate property prediction model, performs latent-space optimization to propose compositions, and then synthesizes and measures five alloys experimentally. The standout result (Ni49.8Ti26.4Hf18.6Zr5.2 alloy reaching 404 °C transformation temperature, 9.9 J/cm³ work output, etc.) is reported from direct characterization, not from the model's output being re-fed as input. No equation or section shows a fitted parameter being relabeled as a prediction, nor does any load-bearing claim rest solely on self-citation of an unverified uniqueness theorem. Self-citations to prior GAN or property-model work exist but are not circular because the present manuscript supplies independent experimental falsification outside the training distribution. The derivation chain therefore remains self-contained once the experimental step is included.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach depends on standard assumptions of generative models and property predictors without introducing new free parameters, axioms beyond domain norms, or invented entities in the reported abstract.

axioms (1)
  • domain assumption Pretrained GAN captures relevant distribution of alloy compositions and processing parameters sufficiently for inversion.
    Invoked to justify latent space optimization leading to valid candidates.

pith-pipeline@v0.9.0 · 5790 in / 1192 out tokens · 60600 ms · 2026-05-19T00:06:13.261178+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

94 extracted references · 94 canonical work pages · 4 internal anchors

  1. [1]

    Otsuka, X

    K. Otsuka, X. Ren, Physical metallurgy of Ti-Ni-based shape memory alloys, Progress in Materials Science 50 (5) (2005) 511– 678

    work page 2005

  2. [2]

    M. H. Elahinia, M. Hashemi, M. Tabesh, S. B. Bhaduri, Manu- facturing and processing of NiTi implants: A review, Progress in Materials Science 57 (5) (2012) 911–946

    work page 2012

  3. [3]

    Mohd Jani, M

    J. Mohd Jani, M. Leary, A. Subic, M. A. Gibson, A review of shape memory alloy research, applications and opportunities, Materials & Design (1980-2015) 56 (2014) 1078–1113

    work page 1980

  4. [4]

    Saedi, E

    S. Saedi, E. Acar, H. Raji, S. E. Saghaian, M. Mirsayar, Energy damping in shape memory alloys: A review, Journal of Alloys and Compounds 956 (2023) 170286

    work page 2023

  5. [5]

    J. Ma, I. Karaman, R. D. Noebe, High temperature shape memory alloys, International Materials Reviews 55 (5) (2013) 257–315

    work page 2013

  6. [6]

    A. V . Shuitcev, Q. Z. Li, M. G. Khomutov, L. Li, Y . X. Tong, Ultra-high temperature shape memory in high-Hf content NiTiHf alloys, Journal of Materials Science & Technology 209 (2025) 124–127

    work page 2025

  7. [7]

    Y . Tong, A. Shuitcev, Y . Zheng, Recent Development of TiNi- Based Shape Memory Alloys with High Cycle Stability and High Transformation Temperature, Advanced Engineering Materials 22 (4) (2020). 10

    work page 2020

  8. [8]

    Amadi, M

    A. Amadi, M. Mohyaldinn, S. Ridha, V . Ola, Advancing engi- neering frontiers with NiTi shape memory alloys: A multifaceted review of properties, fabrication, and application potentials, Jour- nal of Alloys and Compounds 976 (2024) 173227

    work page 2024

  9. [9]

    C. Oses, C. Toher, S. Curtarolo, High-entropy ceramics, Nature Reviews Materials 5 (4) (2020) 295–309

    work page 2020

  10. [10]

    G. L. W. Hart, T. Mueller, C. Toher, S. Curtarolo, Machine learning for alloys, Nature Reviews Materials 6 (8) (2021) 730– 755

    work page 2021

  11. [11]

    J. M. Rickman, T. Lookman, S. V . Kalinin, Materials informatics: From the atomic-level to the continuum, Acta Materialia 168 (2019) 473–510

    work page 2019

  12. [12]

    Y . Xian, P. Dang, Y . Tian, X. Jiang, Y . Zhou, X. Ding, J. Sun, T. Lookman, D. Xue, Compositional design of multicomponent alloys using reinforcement learning, Acta Materialia 274 (2024) 120017

    work page 2024

  13. [13]

    S. H. Zadeh, T. D. Brown, X. Qian, I. Karaman, R. Arroyave, A composition-based predictive model for the transformation strain of NiTi shape memory alloys, Acta Materialia 289 (2025) 120861

    work page 2025

  14. [14]

    C. Li, Q. Liang, Y . Zhou, D. Xue, A knowledge-based materials descriptor for compositional dependence of phase transformation in NiTi shape memory alloys, Materials Genome Engineering Advances 3 (1) (2025) e72

    work page 2025

  15. [15]

    A. A. Catal, E. Bedir, R. Yilmaz, D. Canadinc, Design of a NiTiHf shape memory alloy with an austenite finish temperature beyond 400 °C utilizing artificial intelligence, Journal of Alloys and Compounds 904 (2022) 164135

    work page 2022

  16. [16]

    Hossein Zadeh, C

    S. Hossein Zadeh, C. Cakirhan, D. Khatamsaz, J. Broucek, T. D. Brown, X. Qian, I. Karaman, R. Arroyave, Data-driven study of composition-dependent phase compatibility in NiTi shape memory alloys, Materials & Design 244 (2024) 113096

    work page 2024

  17. [17]

    E. L. Pang, G. B. Olson, C. A. Schuh, Low-hysteresis shape- memory ceramics designed by multimode modelling, Nature 610 (7932) (2022) 491–495

    work page 2022

  18. [18]

    D. Xue, D. Xue, R. Yuan, Y . Zhou, P. V . Balachandran, X. Ding, J. Sun, T. Lookman, An informatics approach to transformation temperatures of NiTi-based shape memory alloys, Acta Materi- alia 125 (2017) 532–541

    work page 2017

  19. [19]

    Broucek, D

    J. Broucek, D. Khatamsaz, C. Cakirhan, S. H. Zadeh, M. Fan, G. Vazquez, K. C. Atli, X. Qian, R. Arroyave, I. Karaman, De- sign of high-temperature NiTiCuHf shape memory alloys with minimum thermal hysteresis using Bayesian optimization, Acta Materialia 286 (2025) 120651

    work page 2025

  20. [20]

    Y . Tian, B. Hu, P. Dang, J. Pang, Y . Zhou, D. Xue, Noise-Aware Active Learning to Develop High-Temperature Shape Memory Alloys with Large Latent Heat, Advanced Science (2024)

    work page 2024

  21. [21]

    Y . Wang, P. Dang, Y . Xian, Y . Zhou, X. Ding, J. Sun, D. Xue, Uncertainty-aware multi-objective optimization for high work output and low hysteresis in TiNiCuHfCo shape memory alloys, Journal of Materials Science & Technology (2025)

    work page 2025

  22. [22]

    D. Xue, P. V . Balachandran, J. Hogden, J. Theiler, D. Xue, T. Lookman, Accelerated search for materials with targeted prop- erties by adaptive design, Nat Commun 7 (2016) 11241

    work page 2016

  23. [23]

    Y . Tian, T. Li, J. Pang, Y . Zhou, D. Xue, X. Ding, T. Lookman, Materials design with target-oriented Bayesian optimization, npj Computational Materials 11 (1) (2025) 209

    work page 2025

  24. [24]

    Canadinc, E

    D. Canadinc, E. J. Breitbach, A. A. Catal, Characterization of Two Novel NiTiHf Shape Memory Alloys Designed by Machine Learning Utilizing Novel Experimental Techniques, Shape Mem- ory and Superelasticity (2025)

    work page 2025

  25. [25]

    B. Shi, T. Lookman, D. Xue, Multi-objective optimization and its application in materials science, Materials Genome Engineering Advances 1 (2) (2023) e14

    work page 2023

  26. [26]

    L. Ding, Y . Zhou, Y . Xu, P. Dang, X. Ding, J. Sun, T. Lookman, D. Xue, Learning from superelasticity data to search for Ti-Ni alloys with large elastocaloric effect, Acta Materialia 218 (2021) 117200

    work page 2021

  27. [27]

    C. Zeni, R. Pinsler, D. Z ¨ugner, A. Fowler, M. Horton, X. Fu, Z. Wang, A. Shysheya, J. Crabb´e, S. Ueda, R. Sordillo, L. Sun, J. Smith, B. Nguyen, H. Schulz, S. Lewis, C.-W. Huang, Z. Lu, Y . Zhou, H. Yang, H. Hao, J. Li, C. Yang, W. Li, R. Tomioka, T. Xie, A generative model for inorganic materials design, Nature 639 (8055) (2025) 624–632

    work page 2025

  28. [28]

    J. Wu, L. Torresi, M. Hu, P. Reiser, J. Zhang, J. S. Rocha-Ortiz, L. Wang, Z. Xie, K. Zhang, B.-w. Park, A. Barabash, Y . Zhao, J. Luo, Y . Wang, L. L¨uer, L.-L. Deng, J. A. Hauch, D. M. Guldi, M. E. P´erez-Ojeda, S. I. Seok, P. Friederich, C. J. Brabec, Inverse design workflow discovers hole-transport materials tailored for perovskite solar cells, Scienc...

    work page 2024

  29. [29]

    Sanchez-Lengeling, A

    B. Sanchez-Lengeling, A. Aspuru-Guzik, Inverse molecular de- sign using machine learning: Generative models for matter engi- neering, Science 361 (6400) (2018) 360–365

    work page 2018

  30. [30]

    S. Lu, Q. Zhou, X. Chen, Z. Song, J. Wang, Inverse design with deep generative models: next step in materials discovery, Natl Sci Rev 9 (8) (2022) nwac111

    work page 2022

  31. [31]

    Cleeton, L

    C. Cleeton, L. Sarkisov, Inverse design of metal-organic frame- works using deep dreaming approaches, Nature Communications 16 (1) (2025) 4806

    work page 2025

  32. [32]

    D. P. Kingma, M. Welling, Auto-Encoding Variational Bayes (De- cember 01, 2013 2013). doi:10.48550/arXiv.1312.6114

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1312.6114 2013

  33. [33]

    I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde- Farley, S. Ozair, A. Courville, Y . Bengio, Generative Adversarial Networks (June 01, 2014 2014). doi:10.48550/arXiv.1406. 2661

    work page doi:10.48550/arxiv.1406 2014

  34. [34]

    Rao, P.-Y

    Z. Rao, P.-Y . Tung, R. Xie, Y . Wei, H. Zhang, A. Ferrari, T. Klaver, F. K¨ormann, P. T. Sukumar, A. Kwiatkowski da Silva, et al., Machine learning–enabled high-entropy alloy discovery, Science 378 (6615) (2022) 78–85

    work page 2022

  35. [35]

    A. Roy, A. Hussain, P. Sharma, G. Balasubramanian, M. F. N. Taufique, R. Devanathan, P. Singh, D. D. Johnson, Rapid dis- covery of high hardness multi-principal-element alloys using a generative adversarial network model, Acta Materialia 257 (2023) 119177

    work page 2023

  36. [36]

    Lee, C.-Y

    C.-Y . Lee, C.-Y . Jui, A.-C. Yeh, Y .-J. Chang, W.-J. Lee, Inverse design of high entropy alloys using a deep interpretable scheme for materials attribution analysis, Journal of Alloys and Com- pounds 976 (2024) 173144

    work page 2024

  37. [37]

    Z. Q. Chen, Y . H. Shang, X. D. Liu, Y . Yang, Accelerated dis- covery of eutectic compositionally complex alloys by generative machine learning, npj Computational Materials 10 (1) (2024) 204

    work page 2024

  38. [38]

    X. Qin, Q. Wang, X. Zhao, S. Xia, L. Wang, Y . Zhang, C. He, D. Chen, B. Jiang, PCS: Property-composition-structure chain in Mg-Nd alloys through integrating sigmoid fitting and conditional generative adversarial network modeling, Scripta Materialia 265 (2025) 116762

    work page 2025

  39. [39]

    Z. Cao, Q. Liu, Q. Liu, X. Yu, J. J. Kruzic, X. Li, A machine learning method to quantitatively predict alpha phase morphol- ogy in additively manufactured Ti-6Al-4V, npj Computational Materials 9 (1) (2023) 195

    work page 2023

  40. [40]

    B. Peng, Y . Wei, Y . Qin, J. Dai, Y . Li, A. Liu, Y . Tian, L. Han, Y . Zheng, P. Wen, Machine learning-enabled constrained multi- objective design of architected materials, Nature Communica- tions 14 (1) (2023) 6630

    work page 2023

  41. [41]

    Z. Li, N. Birbilis, NSGAN: a non-dominant sorting optimisation- based generative adversarial design framework for alloy discov- ery, npj Computational Materials 10 (1) (2024) 112

    work page 2024

  42. [42]

    J. Cai, M. Han, X. Yan, Y . Chen, D. Li, K. Zhao, D. Zhang, K. Hu, H. H. Sua, H. K. Jun, K. Xie, G. Liu, X. Liu, S. Liu, A 11 process-synergistic active learning framework for high-strength Al-Si alloys design, npj Computational Materials 11 (1) (2025) 228

    work page 2025

  43. [43]

    Z. Zhou, Y . Shang, X. Liu, Y . Yang, A generative deep learning framework for inverse design of compositionally complex bulk metallic glasses, npj Computational Materials 9 (1) (2023) 15

    work page 2023

  44. [44]

    Attari, D

    V . Attari, D. Khatamsaz, D. Allaire, R. Arroyave, Towards in- verse microstructure-centered materials design using generative phase-field modeling and deep variational autoencoders, Acta Materialia 259 (2023) 119204

    work page 2023

  45. [45]

    X. Ma, Y . Zhang, C. Wang, M. Wang, M. Huang, W. Xu, Structure-to-process modeling drives experimentally validated unified dual-phase steel, Acta Materialia 295 (2025) 121167

    work page 2025

  46. [46]

    Zang, P.-S

    Y . Zang, P.-S. Koutsourelakis, PSP-GEN: Stochastic inversion of the Process–Structure–Property chain in materials design through deep, generative probabilistic modeling, Acta Materialia 284 (2025) 120600

    work page 2025

  47. [47]

    K. Y . Li, L. C. Liu, L. L. Shao, J. Zhou, H. B. Ke, M. Z. Li, W. H. Wang, Generative model-based inverse design of Fe-based metal- lic glasses with high saturation magnetic flux density, Journal of Alloys and Compounds 1010 (2025) 178325

    work page 2025

  48. [48]

    Abu-Mualla, E

    M. Abu-Mualla, E. Crabtree, F. Michael, Y . Pan, J. Huang, In- verse Design of Alloys via Generative Algorithms: Optimization and Diffusion within Learned Latent Space, Advanced Intelligent Discovery (2025)

    work page 2025

  49. [49]

    P. Dang, J. Hu, Y . Xian, C. Li, Y . Zhou, X. Ding, J. Sun, D. Xue, Elastocaloric thermal battery: Ultrahigh heat-storage capacity based on generative learning-designed phase-change alloys, Ad- vanced Materials 37 (8) (2025) 2412198

    work page 2025

  50. [50]

    Fast Adaptation in Generative Models with Generative Matching Networks

    S. Bartunov, D. P. Vetrov, Fast Adaptation in Generative Models with Generative Matching Networks (December 01, 2016 2016). doi:10.48550/arXiv.1612.02192

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1612.02192 2016

  51. [51]

    S. K. Hung, J. Q. Gan, Small facial image dataset augmentation using conditional GANs based on incomplete edge feature input, PeerJ Comput Sci 7 (2021) e760

    work page 2021

  52. [52]

    2021 , journal =

    A. Andonian, S. Osmany, A. Cui, Y . Park, A. Jahanian, A. Tor- ralba, D. Bau, Paint by Word (March 01, 2021 2021). doi: 10.48550/arXiv.2103.10951

    work page doi:10.48550/arxiv.2103.10951 2021

  53. [53]

    Improved Training of Wasserstein GANs

    I. Gulrajani, F. Ahmed, M. Arjovsky, V . Dumoulin, A. Courville, Improved Training of Wasserstein GANs (March 01, 2017 2017). doi:10.48550/arXiv.1704.00028

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1704.00028 2017

  54. [54]

    Wasserstein GAN

    M. Arjovsky, S. Chintala, L. Bottou, Wasserstein GAN (January 01, 2017 2017). doi:10.48550/arXiv.1701.07875

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1701.07875 2017

  55. [55]

    P. Dang, L. Zhang, Y . Zhou, C. Li, X. Ding, J. Sun, D. Xue, Nanostructured Ti–Ni–Co Alloys Showing Shape-Memory Ac- tuation of Large Work Output at Low Temperature, Advanced Engineering Materials 26 (4) (2023) 2301438

    work page 2023

  56. [56]

    Babacan, M

    N. Babacan, M. Bilal, C. Hayrettin, J. Liu, O. Benafan, I. Kara- man, Effects of cold and warm rolling on the shape memory response of Ni50Ti30Hf20 high-temperature shape memory alloy, Acta Materialia 157 (2018) 228–244

    work page 2018

  57. [57]

    J. Dong, A. Demblon, T. Umale, D. Zhao, G. Valentino, I. Kara- man, K. Y . Xie, Property and microstructure of Ni50.3Ti29.7Hf20 high-temperature shape memory alloys with different aging con- ditions, Acta Materialia 265 (2024) 119642

    work page 2024

  58. [58]

    Baradari, N

    S. Baradari, N. Resnina, S. Belyaev, M. Nili-Ahmadabadi, Martensitic phase transformation and shape memory properties of the as-cast NiCuTiHf and NiCuTiHfZr alloys, Journal of Al- loys and Compounds 888 (2021) 161534

    work page 2021

  59. [59]

    Chen, N.-H

    C.-H. Chen, N.-H. Lu, Improved functional stability of Ti-rich TiNi shape memory ribbon prepared by melt-spinning, Journal of Alloys and Compounds 819 (2020) 152988

    work page 2020

  60. [60]

    H. Li, X. Meng, W. Cai, Shape memory behaviors in a Ti50Ni33.5Cu12.5Pd4 alloy with near-zero thermal hysteresis, Jour- nal of Alloys and Compounds 765 (2018) 166–170

    work page 2018

  61. [61]

    G. S. Bigelow, A. Garg, O. Benafan, R. D. Noebe, S. A. Padula, D. J. Gaydosh, Development and testing of a Ni50.5Ti27.2Hf22.3 high temperature shape memory alloy, Materialia 21 (2022) 101297

    work page 2022

  62. [62]

    S. M. Saghaian, H. E. Karaca, M. Souri, A. S. Turabi, R. D. Noebe, Tensile shape memory behavior of Ni50.3Ti29.7Hf20 high temperature shape memory alloys, Materials & Design 101 (2016) 340–345

    work page 2016

  63. [63]

    Huang, H

    K. Huang, H. Yin, M. Li, Q. Sun, Grain size dependence of stress-assisted two-way memory effect in Ti-50.04 at.% Ni shape memory alloy, Materials Science and Engineering: A 856 (2022) 143872

    work page 2022

  64. [64]

    Imahashi, M

    M. Imahashi, M. I. Khan, H. Y . Kim, S. Miyazaki, The effect of Pd content on microstructure and shape-memory properties of Ti–Ni–Pd–Cu alloys, Materials Science and Engineering: A 602 (2014) 19–24

    work page 2014

  65. [65]

    S. u. Rehman, M. Khan, A. Nusair Khan, L. Ali, S. Zaman, M. Waseem, L. Ali, S. H. I. Jaffery, Transformation behavior and shape memory properties of Ti50Ni15Pd25Cu10 high temperature shape memory alloy at various aging temperatures, Materials Science and Engineering: A 619 (2014) 171–179

    work page 2014

  66. [66]

    Tasaki, Y

    W. Tasaki, Y . Arai, S. Miyazaki, H. Y . Kim, Develop- ment of Ni–Ti–Zr–Hf–(Nb, Ta) Multi-Principal Element High- Temperature Shape Memory Alloys with High Cold Workability, MATERIALS TRANSACTIONS 64 (10) (2023) 2457–2465

    work page 2023

  67. [67]

    Chen, Y .-J

    C.-H. Chen, Y .-J. Chen, Shape memory characteristics of (TiZrHf)50Ni25Co10Cu15 high entropy shape memory alloy, Scripta Materialia 162 (2019) 185–189

    work page 2019

  68. [68]

    J. Pang, Y . Xu, J. Tian, Y . Zhou, D. Xue, X. Ding, J. Sun, Effect of Ti/Ni and Hf/Zr ratio on the martensitic transformation behavior and shape memory effect of TiNiHfZr alloys, Materials Science and Engineering: A 807 (2021) 140850

    work page 2021

  69. [69]

    J. Pang, P. Dang, J. Tian, L. Zhang, Y . Zhou, X. Ding, J. Sun, D. Xue, Effect of Cu content on martensitic transformation and shape memory behavior in Ti31.5Hf15Zr5Ni48.5−xCux alloys, Jour- nal of Materials Science 59 (24) (2024) 11096–11109

    work page 2024

  70. [70]

    E. Acar, G. P. Toker, H. Kurkcu, H. E. Karaca, High tempera- ture shape memory behavior of Ni47.3Ti29.7Hf20Pd3 alloys, Inter- metallics 111 (2019) 106518

    work page 2019

  71. [71]

    G. P. Toker, S. Saedi, E. Acar, O. Ozbulut, H. E. Karaca, Shape memory behavior of Ni 40.3Ti39.7Hf15Pd5 and Ni40.3Ti44.7Hf10Pd5 alloys, Journal of Alloys and Compounds 763 (2018) 1012–1017

    work page 2018

  72. [72]

    Saedi, G

    S. Saedi, G. P. Toker, H. Raji, E. Acar, F. Watanabe, H. Karaca, Shape Memory Effect in Quaternary NiTiHfPd Shape Memory Alloys beyond 200 ◦C, Journal of Materials Engineering and Performance 31 (8) 6689–6696

    work page

  73. [73]

    Canadinc, W

    D. Canadinc, W. Trehern, J. Ma, I. Karaman, F. Sun, Z. Chaudhry, Ultra-high temperature multi-component shape memory alloys, Scripta Materialia 158 (2019) 83–87

    work page 2019

  74. [74]

    H. Y . Kim, T. Jinguu, T.-h. Nam, S. Miyazaki, Cold work- ability and shape memory properties of novel Ti–Ni–Hf–Nb high-temperature shape memory alloys, Scripta Materialia 65 (9) (2011) 846–849

    work page 2011

  75. [75]

    Evirgen, I

    A. Evirgen, I. Karaman, R. Santamarta, J. Pons, R. D. Noebe, Microstructural characterization and shape memory characteris- tics of the Ni50.3Ti34.7Hf15 shape memory alloy, Acta Materialia 83 (2015) 48–60

    work page 2015

  76. [76]

    Evirgen, I

    A. Evirgen, I. Karaman, R. D. Noebe, R. Santamarta, J. Pons, Ef- fect of precipitation on the microstructure and the shape memory response of the Ni50.3Ti29.7Zr20 high temperature shape memory alloy, Scripta Materialia 69 (5) (2013) 354–357

    work page 2013

  77. [77]

    H. E. Karaca, S. M. Saghaian, G. Ded, H. Tobe, B. Basaran, H. J. Maier, R. D. Noebe, Y . I. Chumlyakov, Effects of nanoprecipita- tion on the shape memory and material properties of an Ni-rich 12 NiTiHf high temperature shape memory alloy, Acta Materialia 61 (19) (2013) 7422–7431

    work page 2013

  78. [78]

    Karakoc, C

    O. Karakoc, C. Hayrettin, M. Bass, S. J. Wang, D. Canadinc, J. H. Mabe, D. C. Lagoudas, I. Karaman, Effects of upper cycle temperature on the actuation fatigue response of NiTiHf high temperature shape memory alloys, Acta Materialia 138 (2017) 185–197

    work page 2017

  79. [79]

    Demblon, J

    A. Demblon, J. H. Mabe, I. Karaman, Compositional effects on strain-controlled actuation fatigue of NiTiHf high temperature shape memory alloys, Scripta Materialia 242 (2024) 115904

    work page 2024

  80. [80]

    H. Chen, F. Xiao, X. Liang, Z. Li, Z. Li, X. Jin, T. Fukuda, Giant elastocaloric effect with wide temperature window in an Al-doped nanocrystalline Ti–Ni–Cu shape memory alloy, Acta Materialia 177 (2019) 169–177

    work page 2019

Showing first 80 references.