Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

Graph-to-SFILES: Control structure prediction from process topologies using generative artificial intelligence

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Representing a process flowsheet as a graph instead of a text sequence makes a generative model far better at predicting control structures when training data are scarce.

desk verdict A useful, honest proof-of-concept for graph-based control-structure prediction, but the headline small-data advantage over sequence models rests on an undertuned baseline that should be re-run before the claim is trusted. read the letter →

arxiv 2412.00508 v2 pith:S25LQ22F submitted 2024-11-30 cs.LG cs.AIcs.CE

Lukas Schulze Balhorn , Kevin Degens , Artur M. Schweidtmann This is my paper · ORCID
classification cs.LGcs.AIcs.CE
keywords controlstructurepredictiongraph-to-sequenceprocessgraphSFILES2.0generativeartificialintelligenceneuralnetworktransformerdecoderflowdiagram
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

Control structure design—choosing the controllers that keep a chemical process inside its operating window—is a manual step in P&ID development. This paper argues that a generative model can learn this step from flowsheet topology alone if the topology is presented as a graph rather than a text string. The proposed Graph-to-SFILES model encodes a process flow diagram as a permutation-invariant graph and decodes it into an SFILES 2.0 string that includes the added controllers. The central result is that on a small training set of 1,000 flowsheets, the graph-based model reaches 28.4% top-5 accuracy while a comparable sequence-based model reaches only 0.9%. The paper also contributes a graph encoder architecture, the Combined model, that outperforms three literature encoders, and it argues this data efficiency matters because detailed industrial process data are scarce.

What carries the argument

The carrying mechanism is the Graph-to-SFILES model: a graph encoder built from graph neural network layers that message-pass over nodes (unit operations, valves, controllers) and edges (streams, signals), followed by an autoregressive transformer decoder that emits the control-extended flowsheet as an SFILES 2.0 string under beam search. SFILES 2.0 is a SMILES-like text notation that encodes flowsheet topology and control elements, including controller tags such as {PC} and signal-line markers, with a canonical form for scoring. The paper's new encoder, the Combined model, merges the dynamic attention of GATv2 with the Laplacian positional encoding and residual connections of the Graph Transformer; the permutation invariance of the graph input is what removes the need for sequence-order augmentation and is credited for the small-data improvement.

What would settle it

A dataset in which the same PFD topology appears with different stream parameters and different ground-truth controllers would settle it: if Graph-to-SFILES cannot distinguish those pairs, topology alone is insufficient, and the claimed small-data advantage would not transfer to real P&IDs.

Watch

Extended reading notes

Core claim

The paper's central claim is that control-structure prediction from process topologies is better posed as a graph-to-sequence translation than as a sequence-to-sequence translation when training data are scarce, and that a graph neural network encoder with positional encoding is the best encoder for this task. On 1,000 training flowsheets, Graph-to-SFILES with the proposed Combined encoder reaches a top-1 CEF accuracy of 15.2% and top-5 of 28.4%, versus 0.3% and 0.9% for the sequence baseline; with 10,000 flowsheets it reaches 73.2% top-5. The advantage reverses at 100,000 flowsheets, where the sequence model scores 89.2% top-5 against 79.0%, so the paper's conclusion is scoped to the small-data regime.

Load-bearing premise

The load-bearing premise is that a control structure can be learned from flowsheet topology alone: the synthetic training pairs contain no equipment, operation, or stream parameters, so the model can only pick up the mapping from topology plus whatever bias the synthetic data carry.

Editorial extensions

If this is right

  • A practical support tool for P&ID development becomes plausible: an engineer draws the PFD topology and the model proposes the top five control-structure strings, with the third-best beam matching the ground truth in the paper's illustrative example.
  • For teams with small labeled datasets, which the paper argues is the realistic situation for detailed process data, graph representation should be preferred over sequence representation for this task.
  • Adding equipment, operation, and stream parameters as node and edge features is a direct extension that the paper identifies as the open question for industrial transfer.
  • Forcing the decoder to reproduce the input PFD topology in the output SFILES could remove the largest source of error, since PFD reconstruction accuracy is always below 100%.
  • At very large data scales the sequence baseline remains competitive or superior, so representation choice should depend on data availability.

Reading between the lines

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

  • If real control-structure design depends on pressures, temperatures, or compositions, the 28.4% small-data gain reflects the synthetic dataset's label distribution as much as the graph representation; a dataset with parameter-dependent labels would separate the two.
  • The permutation-invariance advantage should grow on real PFDs with arbitrary node ordering, because sequence models would need canonicalization or augmentation that graph models can skip.
  • Because the output is text, the same encoder-decoder could be pointed at other P&ID annotations, such as instrument lists, alarm/HAZOP text, or flowsheet autocorrection, without changing the architecture.
  • A decoder constrained to regenerate the input graph (for example, by copying node identities) would likely close much of the PFD reconstruction gap and, with it, raise CEF accuracy.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper proposes Graph-to-SFILES, a graph-to-sequence model that takes a process flow diagram (PFD) represented as a graph and generates a control-extended flowsheet (CEF) in SFILES 2.0 notation. Four graph encoder architectures are compared, including a new 'Combined' encoder that merges GATv2-style attention with Graph Transformer positional encodings and residual connections. The models are trained on synthetic datasets of 1k, 10k, and 100k PFD–CEF pairs. On 10k training flowsheets, the best model reaches 73.2% top-5 CEF accuracy; on 1k flowsheets, the graph-based model improves top-5 CEF accuracy from 0.9% (sequence baseline) to 28.4%, while on 100k flowsheets the sequence baseline performs better. The authors argue that graph representations are more data-efficient in small-data regimes and position the work as a proof-of-concept for AI-assisted control structure design.

Significance. If the empirical claims are fully supported, the paper is a meaningful contribution to AI-assisted process design. It provides a concrete comparison of graph and sequence representations for a generative task in chemical engineering, and the finding that graphs help in the small-data regime is practically relevant because real process datasets are scarce. The paper has notable strengths: it reports multiple seeds with standard deviations, evaluates on an independent test set, and honestly states limitations in Section 5.4 (no process parameters, synthetic data, purely data-driven). The proposal of a new encoder architecture is a reasonable incremental contribution. However, the central graph-versus-sequence comparison currently rests on a baseline that is not tuned under the same protocol as the proposed models, which weakens the headline quantitative claim.

major comments (3)
  1. [Section 5.2, Table 2] The comparison between the Graph-to-SFILES model and the sequence-to-sequence baseline is not a controlled experiment. Section 5.1 states that hyperparameters for the four graph encoders were optimized via grid search on the 10k validation set, but no evidence is provided that the sequence baseline from Hirtreiter et al. (2023) was re-tuned on the 1k, 10k, or 100k datasets, nor that it was trained with the same number of seeds or epochs. Since the headline claim of the abstract (0.9% to 28.4% top-5 improvement) depends on this comparison, the authors should either re-run the sequence baseline under the same optimization and training protocol or justify why the prior hyperparameters are appropriate across all dataset sizes. Without this, the small-data advantage attributed to graph inputs could be partly a training-budget or hyperparameter artifact.
  2. [Section 5.3, point (v), and Section 5.4] The CEF accuracy metric counts a prediction as correct only if the generated SFILES matches the ground truth exactly. In the illustrative example, the authors themselves note that several of the model's top-5 predictions (e.g., TC instead of TI, FFC instead of FC) are 'valid process configurations' that appear in the training data. Because control structure design is non-unique, exact-match accuracy likely underestimates the model's practical performance, and the absolute accuracy numbers reported in the abstract and Table 2 may be misleading as measures of producing valid control structures. The authors should either introduce an evaluation metric that accepts multiple valid solutions or explicitly frame the reported accuracies as conservative lower bounds and discuss how this affects the headline results.
  3. [Section 5.1, Table 2] The claim that the proposed Combined model 'performs best' among the graph encoders is not strongly supported at the 100k scale. For top-1 CEF accuracy, the Combined model achieves 46.1% ± 0.7% while GraphConv achieves 45.9% ± 0.4%, a difference well within the reported standard deviations. The claim is better supported at 1k and 10k, where the gaps are larger. The paper should either add a statistical significance test (e.g., paired tests across seeds) or temper the claim to state that the Combined model is best on the smaller datasets and competitive with GraphConv at 100k.
minor comments (5)
  1. [Section 1] The statement that 'no graph-based, generative AI method for process development exists yet' is strong; given the breadth of prior work cited, it would be safer to say 'to the best of our knowledge' and perhaps cite recent graph-based process synthesis work more explicitly.
  2. [Section 5.3] The illustrative example selects the model sample with the highest top-1 CEF accuracy among the three trained samples. This is a selection of the best performer and is not representative of average model behavior; the paper should acknowledge this explicitly when presenting the example.
  3. [Table 1] The edge attribute encoding is described as a three-dimensional vector with values 0, 1, 2 for 'Unspecified', 'Top', 'Bottom', etc. It would help to state explicitly whether these are ordinal or categorical encodings, since the interpretation affects the GNN's ability to use them.
  4. [Table 4] The hyperparameter table lists 'Heads m' as 8 for GATv2 but leaves dashes for the other three architectures; it should clarify that the default number of heads is used for the other encoders, or list the actual value.
  5. [Section 5.4] The discussion of data efficiency would be strengthened by a quantitative analysis, such as learning curves or a direct comparison of the scaling behavior of the graph and sequence models across dataset sizes, rather than only a qualitative statement.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reported accuracies are empirical supervised-learning measurements; self-cited inputs (dataset, tokenizer, baseline) are not proofs of the conclusion, and no predicted quantity is defined in terms of the fitting procedure.

full rationale

The paper's central result (Table 2: Combined model top-5 CEF accuracy 28.4% vs 0.9% on 1,000 flowsheets) is a held-out exact-match accuracy comparison on an independent test set, not a quantity that follows by construction from the model definition. The graph encoder and transformer decoder are standard components; the SFILES 2.0 representation and tokenizer are operational encodings, not assumptions that entail the accuracy gap. The sequence baseline values are taken from Hirtreiter et al. (2023), a self-citation, but those numbers are externally reported empirical results and are not re-fit or redefined in the present paper; the possible asymmetry in hyperparameter tuning is a fairness/correctness risk, not a circular reduction. The paper also explicitly flags its main limitations (Section 5.4: no equipment, operation, or stream parameters; synthetic training data; PFD reconstruction errors), which affect applicability and external validity rather than making the in-benchmark result circular. No equation or metric is self-definitional: CEF accuracy, PFD reconstruction, and BLEU are all computed by comparing generated SFILES to ground-truth strings after canonicalization, and none of these metrics is used as a training target or fitted parameter. No 'uniqueness theorem' is invoked to force model choice; the canonical SFILES rule is cited for evaluation only. Therefore no specific circular step can be exhibited.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The model makes no mathematical derivation claims, so the ledger captures the hand-chosen inputs to the empirical pipeline: grid-searched hyperparameters, a hand-designed feature encoding, and the synthetic data generation assumptions. The key unstated cost is the data generation heuristics, which are not released. No new physical entities are introduced.

free parameters (3)
  • Neural network hyperparameters (batch size, learning rate, encoder/decoder layers, heads, embedding dimension… = Table 4: e.g., Combined model uses batch size 64, learning rate 1e-3, 6 encoder and 6 decoder layers
    Selected by grid search on the 10k validation loss, then applied to the 1k and 100k datasets; all reported accuracies depend on these choices.
  • Node and edge feature encoding scheme = Table 1: outlet position (unspecified/top/bottom), hex inlet and outlet side (unspecified/side 1/side 2)
    Hand-designed encoding of graph attributes chosen by the authors; it defines the only process information available to the model.
  • Beam size in beam search decoding = 5 (implied by top-5 metrics)
    The decoder returns top-k predictions; the paper reports top-5, but the beam width itself is not listed as a tuned hyperparameter.
assumptions (4)
  • domain assumption SFILES 2.0 to graph and graph to SFILES conversions are faithful and canonicalization is correct.
    The training pipeline and all metrics depend on the open-source SFILES 2.0 repository (Section 4) for lossless conversion and canonical forms; errors here would corrupt both training labels and evaluation.
  • domain assumption Synthetic flowsheets generated from literature patterns by engineering heuristics are a valid proxy for real process topologies.
    Section 4 states the schematics 'do not represent an actual chemical process but are similar regarding their structure'; the validity of all accuracy claims depends on this proxy.
  • domain assumption Topology-only PFDs contain enough information to infer a suitable control structure.
    The input graphs contain no equipment, operation, or stream parameters (Sections 2.1 and 5.4). The authors acknowledge the model must rely on dataset bias, so this assumption is load-bearing for generalization.
  • standard math Standard supervised learning assumptions hold: independent train/validation/test splits, teacher forcing, and cross-entropy training.
    Section 3.4 describes teacher-forced training and Section 5.1 reports evaluation on an independent test set; no formal guarantees are given for generalization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Graph-to-SFILES: Control structure prediction from process topologies using generative artificial intelligence." pith.science (2026). https://pith.science/paper/S25LQ22F

@misc{pith2026241200508,
  author       = {Pith},
  title        = {Pith review of: Graph-to-SFILES: Control structure prediction from process topologies using generative artificial intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S25LQ22F}},
  note         = {Machine review of arXiv:2412.00508}
}
read the original abstract

Control structure design is an important but tedious step in P&ID development. Generative artificial intelligence (AI) promises to reduce P&ID development time by supporting engineers. Previous research on generative AI in chemical process design mainly represented processes by sequences. However, graphs offer a promising alternative because of their permutation invariance. We propose the Graph-to-SFILES model, a generative AI method to predict control structures from flowsheet topologies. The Graph-to-SFILES model takes the flowsheet topology as a graph input and returns a control-extended flowsheet as a sequence in the SFILES 2.0 notation. We compare four different graph encoder architectures, one of them being a graph neural network (GNN) proposed in this work. The Graph-to-SFILES model achieves a top-5 accuracy of 73.2% when trained on 10,000 flowsheet topologies. In addition, the proposed GNN performs best among the encoder architectures. Compared to a purely sequence-based approach, the Graph-to-SFILES model improves the top-5 accuracy for a relatively small training dataset of 1,000 flowsheets from 0.9% to 28.4%. However, the sequence-based approach performs better on a large-scale dataset of 100,000 flowsheets. These results highlight the potential of graph-based AI models to accelerate P&ID development in small-data regimes but their effectiveness on industry relevant case studies still needs to be investigated.

Figures

Figures reproduced from arXiv: 2412.00508 by the authors.

Figure 1
Figure 1. Graph representation of flowsheets. 2.1 Process graphs Chemical processes can be viewed as complex networks of unit operations connected through streams. An intuitive choice for the information representation is therefore a graph. We make use of the graph format defined by Vogel et al. (2023b) for PFDs. Here, nodes represent unit operations, control units, raw materials, and products, while directed edges represent … view at source ↗
Figure 2
Figure 2. Overview of Graph-to-SFILES methodology. The PFD is represented as a graph with node and edge features. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Detailed Graph-to-SFILES model consisting of a graph encoder and a sequence decoder. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: GNN layers as a drop-in replacement in the graph encoder. Each GNN layer takes the node features [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Graph representation of a PFD, Input to the graph encoder. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Illustrative example PFD. We employ the Combined model encoder trained on 100,000 flowsheets. From the three model samples presented in 2, we select the sample with the highest top-1 CEF accuracy. The model predicts the following five SFILES 2.0. Here, the five predict…
Figure 7
Figure 7. Figure 7: Control structure prediction for the sample PFD (colored black) where correct predictions are shown in blue [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. FaultExplainer: Leveraging Large Language Models for Interpretable Fault Detection and Diagnosis

    cs.AI 2024-12 conditional novelty 6.0 of 10

    An LLM chatbot grounded in PCA feature contributions gives plausible root-cause explanations for Tennessee Eastman Process faults, but only for faults the PCA detector can see.

Reference graph

Works this paper leans on

48 extracted references · 39 canonical work pages · cited by 1 Pith paper

  1. [1]

    R., Nabilc, T., Loubièreb, V., Valentinc, B., Privata, R., Neveuxb, T., and Commenge, J.-M

    Azevedoa, A. R., Nabilc, T., Loubièreb, V., Valentinc, B., Privata, R., Neveuxb, T., and Commenge, J.-M. (2024). Representation learning for flowsheets: generating structures for process synthesis. In Manenti, F. and Reklaitis, G. V., editors, Book of Abstract of the 34th European Symposium on Computer Aided Process Engineering

    work page 2024

  2. [2]

    S., Caballero, M., and Schweidtmann, A

    Balhorn, L. S., Caballero, M., and Schweidtmann, A. M. (2024). Toward autocorrection of chemical process flowsheets using large language models. In Computer Aided Chemical Engineering , volume 53, pages 3109--3114. Elsevier

    work page 2024

  3. [3]

    S., Gao, Q., Goldstein, D., and Schweidtmann, A

    Balhorn, L. S., Gao, Q., Goldstein, D., and Schweidtmann, A. M. (2022). Flowsheet recognition using deep convolutional neural networks. In Computer Aided Chemical Engineering , volume 49, pages 1567--1572. Elsevier

    work page 2022

  4. [4]

    Beck, D., Haffari, G., and Cohn, T. (2018). Graph-to-sequence learning using gated graph neural networks. arXiv preprint arXiv:1806.09835

    work page Pith review arXiv 2018

  5. [5]

    Bramsiepe, C., Krasberg, N., Fleischer, C., Hohmann, L., Kockmann, N., and Schembecker, G. (2014). Information technologies for innovative process and plant design. Chemie Ingenieur Technik , 86(7):966--981

    work page 2014

  6. [6]

    Brody, S., Alon, U., and Yahav, E. (2021). How attentive are graph attention networks? arXiv preprint arXiv:2105.14491

    arXiv 2021

  7. [7]

    d'Anterroches, L. (2005). Process Flowsheet Generation & Design through a Group Contribution Approach . [CAPEC], Department of Chemical Engineering, Technical University of Denmark

    work page 2005

  8. [8]

    and Kipf, T

    De Cao, N. and Kipf, T. (2018). Molgan: An implicit generative model for small molecular graphs. arXiv preprint arXiv:1805.11973

    arXiv 2018

Show all 48 references
  1. [9]

    Dollar, O., Joshi, N., Pfaendtner, J., and Beck, D. A. (2023). Efficient 3d molecular design with an e (3) invariant transformer vae. The Journal of Physical Chemistry A , 127(37):7844--7852

    2023

  2. [10]

    R., Guo, J., Fu, T., Harris, C., Wang, Y., Duan, C., Li \`o , P., Schwaller, P., and Blundell, T

    Du, Y., Jamasb, A. R., Guo, J., Fu, T., Harris, C., Wang, Y., Duan, C., Li \`o , P., Schwaller, P., and Blundell, T. L. (2024). Machine learning-aided generative molecular design. Nature Machine Intelligence , pages 1--16

    2024

  3. [11]

    Dwivedi, V. P. and Bresson, X. (2020). A generalization of transformer networks to graphs. arXiv preprint arXiv:2012.09699

    2020 arXiv

  4. [12]

    Dzhusupova, R., Banotra, R., Bosch, J., and Olsson, H. H. (2022). Pattern recognition method for detecting engineering errors on technical drawings. In 2022 IEEE World AI IoT Congress (AIIoT) , pages 642--648. IEEE, IEEE

    2022

  5. [13]

    and Schweidtmann, A

    Gao, Q. and Schweidtmann, A. M. (2024). Deep reinforcement learning for process design: Review and perspective. Current Opinion in Chemical Engineering , 44:101012

    2024

  6. [14]

    Hirtreiter, E., Schulze Balhorn, L., and Schweidtmann, A. M. (2023). Toward automatic generation of control structures for process flow diagrams with large language models. AIChE Journal , n/a(n/a):e18259

    2023

  7. [15]

    and Schmidhuber, J

    Hochreiter, S. and Schmidhuber, J. (1997). Long short-term memory. Neural computation , 9(8):1735--1780

    1997

  8. [16]

    Kipf, T. N. and Welling, M. (2016). Variational graph auto-encoders. arXiv preprint arXiv:1611.07308

    2016 arXiv

  9. [17]

    Li, J. (2024). Learning hybrid extraction and distillation using phenomena-based string representation. In Foundations of Computer Aided Process Design (FOCAPD 2024)

    2024

  10. [18]

    L., Tyreus, B

    Luyben, M. L., Tyreus, B. D., and Luyben, W. L. (1997). Plantwide control design procedure. AICHE journal , 43(12):3161--3174

    1997

  11. [19]

    Mann, V., Sales-Cruz, M., Gani, R., and Venkatasubramanian, V. (2024). esfiles: Intelligent process flowsheet synthesis using process knowledge, symbolic ai, and machine learning. Computers & Chemical Engineering , 181:108505

    2024

  12. [20]

    Mao, K., Xiao, X., Xu, T., Rong, Y., Huang, J., and Zhao, P. (2021). Molecular graph enhanced transformer for retrosynthesis prediction. Neurocomputing , 457:193--202

    2021

  13. [21]

    Morari, M., Arkun, Y., and Stephanopoulos, G. (1980). Studies in the synthesis of control structures for chemical processes: Part i: Formulation of the problem. process decomposition and the classification of the control tasks. analysis of the optimizing control structures. AI...

    1980

  14. [22]

    Nabil, T., Le Moullec, Y., and Le Coz, A. (2019). Machine learning based design of a supercritical co2 concentrating solar power plant. In 3rd European supercritical CO2 Conference September 19-20, 2019, Paris, France

    2019

  15. [23]

    Nabil, T., Noaman, M., and Morosuk, T. (2023). Data-driven structural synthesis of supercritical co2 power cycles. Frontiers in Chemical Engineering , 5:1144115

    2023

  16. [24]

    and Stephanopoulos, G

    Ng, C. and Stephanopoulos, G. (1998). Synthesis of control structures for chemical plants. volume 31, pages 17--29. Elsevier

    1998

  17. [25]

    Oeing, J., Brandt, K., Wiedau, M., Tolksdorf, G., Welscher, W., and Kockmann, N. (2023). Graph learning in machine-readable plant topology data. Chemie Ingenieur Technik , 95(7):1049--1060

    2023

  18. [26]

    Oeing, J., Welscher, W., Krink, N., Jansen, L., Henke, F., and Kockmann, N. (2022). Using artificial intelligence to support the drawing of piping and instrumentation diagrams using dexpi standard. Digital Chemical Engineering , 4:100038

    2022

  19. [27]

    Papineni, K., Roukos, S., Ward, T., and Zhu, W.-J. (2002). Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th annual meeting of the Association for Computational Linguistics , pages 311--318

    2002

  20. [28]

    Schweidtmann, A. M. (2024a). Generative artificial intelligence in chemical engineering. Nature Chemical Engineering , 1(3):193--193

    2024

  21. [29]

    Schweidtmann, A. M. (2024b). Mining chemical process information from literature for generative process design: A perspective. In Foundations of Computer Aided Process Design (FOCAPD 2024)

    2024

  22. [30]

    E., Edgar, T

    Seborg, D. E., Edgar, T. F., Mellichamp, D. A., and Doyle III, F. J. (2016). Process dynamics and control . John Wiley & Sons

    2016

  23. [31]

    Shi, Y., Huang, Z., Feng, S., Zhong, H., Wang, W., and Sun, Y. (2020). Masked label prediction: Unified message passing model for semi-supervised classification. arXiv preprint arXiv:2009.03509

    2020 arXiv

  24. [32]

    Song, J. J. and Park, S. W. (1990). Intellite 3: A knowledge based expert system for control structure synthesis. Korean Journal of Chemical Engineering , 7:198--209

    1990

  25. [33]

    Song, L., Zhang, Y., Wang, Z., and Gildea, D. (2018). A graph-to-sequence model for amr-to-text generation. arXiv preprint arXiv:1805.02473

    2018 arXiv

  26. [34]

    Stops, L., Leenhouts, R., Gao, Q., and Schweidtmann, A. M. (2023). Flowsheet generation through hierarchical reinforcement learning and graph neural networks. AIChE Journal , 69(1):e17938

    2023

  27. [35]

    F., Flores, K

    Theisen, M. F., Flores, K. N., Schulze Balhorn, L., and Schweidtmann, A. M. (2023). Digitization of chemical process flow diagrams using deep convolutional neural networks. Digital Chemical Engineering , 6:100072

    2023

  28. [36]

    Toghraei, M. (2019). Piping and instrumentation diagram development . John Wiley & Sons

    2019

  29. [37]

    and Coley, C

    Tu, Z. and Coley, C. W. (2023). Permutation invariant graph-to-sequence model for template-free retrosynthesis and reaction prediction

    2023

  30. [38]

    and De Jager, B

    Van De Wal, M. and De Jager, B. (2001). A review of methods for input/output selection. Automatica , 37(4):487--510

    2001

  31. [39]

    Vaswani, A., Brain, G., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A., and Kaiser, . (2017). Attention is all you need. Advances in Neural Information Processing Systems

    2017

  32. [40]

    Veli c kovi \'c , P., Cucurull, G., Casanova, A., Romero, A., Lio, P., and Bengio, Y. (2017). Graph attention networks. arXiv preprint arXiv:1710.10903

    2017 arXiv

  33. [41]

    S., and Schweidtmann, A

    Vogel, G., Balhorn, L. S., and Schweidtmann, A. M. (2023a). Learning from flowsheets: A generative transformer model for autocompletion of flowsheets. Computers & Chemical Engineering , 171:108162

    2023

  34. [42]

    S., and Schweidtmann, A

    Vogel, G., Hirtreiter, E., Balhorn, L. S., and Schweidtmann, A. M. (2023b). SFILES 2.0: an extended text-based flowsheet representation. Optimization and Engineering

    2023

  35. [43]

    Vogel, G., Sortino, P., and Weber, J. M. (2023c). Graph-to-string variational autoencoder for synthetic polymer design. In AI for Accelerated Materials Design - NeurIPS 2023 Workshop

    2023

  36. [44]

    Wang, J., Zheng, S., Chen, J., and Yang, Y. (2021). Meta learning for low-resource molecular optimization. volume 61, pages 1627--1636. ACS Publications

    2021

  37. [45]

    Weininger, D., Weininger, A., and Weininger, J. L. (1989). Smiles. 2. algorithm for generation of unique smiles notation. Journal of chemical information and computer sciences , 29(2):97--101

    1989

  38. [46]

    Williamson, C. (1989). Computer aided process control systems synthesis using rule-based programming . University of Canterbury. Chemical Engineering

    1989

  39. [47]

    Xu, K., Wu, L., Wang, Z., Feng, Y., Witbrock, M., and Sheinin, V. (2018). Graph2seq: Graph to sequence learning with attention-based neural networks. arXiv preprint arXiv:1804.00823

    2018 arXiv

  40. [48]

    You, J., Ying, R., Ren, X., Hamilton, W., and Leskovec, J. (2018). G raph RNN : Generating realistic graphs with deep auto-regressive models. In Dy, J. and Krause, A., editors, Proceedings of the 35th International Conference on Machine Learning , volume 80 of Proceedings of M...

    2018

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.