Pith. sign in

REVIEW 3 major objections 5 minor 31 references

Transformer-based Deep Learning Model for Joint Routing and Scheduling with Varying Electric Vehicle Numbers

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Pith's one-line read A transformer trained on deterministic EV routing problems predicts the optimal binary decisions well enough to prune the mixed-integer search space, cutting Gurobi runtime by 98.1% on average with 100% feasibility and less than 0.0007%…

desk verdict A legitimate fleet-size-agnostic transformer adaptation for predicting MIP binary variables, but the 98% speedup headline is a selected best-of-four statistic and needs a proper held-out evaluation before I'd trust the number. read the letter →

arxiv 2507.15385 v1 pith:NK7RZOVY submitted 2025-07-21 eess.SY cs.SY

classification eess.SYcs.SY
keywords ElectricVehiclesPowersystemsimulationOptimizationNeuralnetworksVehicleroutingTransformerMixed-integerprogrammingSearchspacepruning
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

This paper tries to show that a transformer-based neural network can make a hard day-ahead optimization problem—scheduling and routing electric vehicles across a power grid and a traffic network—tractable by predicting the optimal binary decisions before a commercial solver starts. The authors argue that fixing the predicted high-confidence binary variables prunes the mixed-integer search space enough that Gurobi reaches near-optimal solutions much faster, and that the model's attention-based design lets it handle any fleet size without retraining. On simulations of an IEEE 33-bus distribution network coupled to the Nguyen-Dupuis transportation network, the assisted solver is claimed to cut runtime by 98.1% on average while keeping every tested instance feasible and losing less than 0.0007% of solution quality. If correct, the method points to a general recipe: use learned structure to shrink the combinatorial core of MIPs, while leaving constraint satisfaction to the solver.

What carries the argument

The load-bearing object is a two-layer transformer encoder whose sequence dimension is repurposed from time to features (bus loads, PV output, and EV job schedules), with time kept as the feature dimension, so that the number of EV-related tokens can vary from instance to instance. After embedding, three feed-forward classification layers with a sigmoid output produce a probability for each binary decision variable, and a thresholding filter derived from the mean prediction probability of each class selects the bits used to fix variables in Gurobi. The same variable-size FFN mechanism lets the model produce predictions for fleets of sizes never seen in training, which is what makes the fleet-size-agnostic claim work.

What would settle it

Solve enough stochastic EVJRS test instances to proven optimality, apply the transformer's thresholded predictions as fixed binary variables, and check whether the true optimal solution of any instance violates a fixed variable; if it does, the pruning step has cut away the optimum and the near-zero quality-loss claim collapses.

Watch

Extended reading notes

Core claim

The central claim is that the optimal binary variables of the EVJRS stochastic mixed-integer program—EV routing choices and charging/discharging statuses—can be predicted well enough by a transformer trained on deterministic instances that the predictions can be used to prune the MIP's search space before Gurobi solves it. The authors report that on a held-out set containing stochastic instances and EV counts unseen during training, the best transformer model achieves 100% feasibility, an average runtime reduction of 98.1% over bare Gurobi, and an average objective loss of only $6.5 \times 10^{-4}\%$ (below 0.0007%), without retraining. They also report that models trained on coarser fleet-size granularity (e.g., EV counts in steps of 15) balance feasibility and solution quality better than finer-grained training. The mechanism is not direct solution: the network emits binary bit probabilities, a thresholding filter keeps only confident bits, and Gurobi completes the problem with the remaining variables.

Load-bearing premise

The load-bearing premise is that binary decisions learned from deterministic instances remain representative of the optimal binary decisions in the full stochastic problem, so that scenario-combined predictions can safely prune the stochastic MIP's search space.

Editorial extensions

If this is right

  • Day-ahead EV coordination with fleets of 20–100 vehicles can be cleared in a fraction of the original solve time, making the approach usable in time-sensitive market settings where bare Gurobi becomes prohibitively slow.
  • A single trained model can be applied to unseen fleet sizes, so operators do not need to retrain whenever the EV fleet changes.
  • Feasibility is guaranteed by the solver rather than approximated by the network, because the network only fixes a subset of binary variables and Gurobi enforces all constraints on the remainder.
  • Coarser training granularity can dominate finer granularity on the metrics that matter: TF15 achieves 100% feasibility with nearly the same runtime reduction as TF20 and much lower objective loss than TF20.
  • The thresholding filter gives a tunable safety valve: raising the confidence threshold retains fewer but safer binary bits, trading a little speed for additional solution-quality insurance.

Reading between the lines

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

  • The deterministic-to-stochastic transfer is the least tested link: the paper trains on deterministic instances and tests on stochastic ones, but does not compare the predicted binary pattern against the true stochastic optimum, so a targeted optimality check on small solvable instances would settle whether the pruning ever discards an optimal solution.
  • The same architecture-pruning recipe should transfer to other MIPs whose binary variables scale with a population—e.g., unit commitment, fleet dispatch, or appliance scheduling—because the transformer's variable-length sequence removes the main obstacle of re-training per problem size.
  • A practical deployment would likely need an online calibration of the threshold, since the confidence filter is currently calibrated on the training data's mean prediction probabilities rather than on an operator-specified worst-case quality bound.
  • Large-fleet extrapolation remains an open question that the authors themselves flag: labels for very large instances are expensive, so a natural test is whether a model trained on small fleets can still prune well when applied to fleets far larger than 100 EVs.
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 a transformer-based deep learning model that predicts optimal binary decisions of a mixed-integer programming formulation of a day-ahead electric vehicle joint routing and scheduling (EVJRS) problem. The predicted binaries are threshold-filtered and used to prune the search space for Gurobi. The model is designed to handle varying EV fleet sizes without retraining, and is trained on deterministic instances while tested on the full stochastic problem. On a test set derived from the IEEE 33-bus distribution network coupled with the Nguyen-Dupuis transportation network, the best transformer variant (TF15) is reported to reduce Gurobi solution time by 98.1% on average, achieve 100% feasibility, and lose less than 0.0007% solution quality, compared with a CNN baseline using padding.

Significance. If the reported results hold, the paper addresses a practically relevant problem: accelerating the solution of stochastic EV routing and scheduling with variable fleet sizes, while preserving feasibility and solution quality. The paper makes its code publicly available, provides a formal MIP formulation with explicit constraints, and compares against a CNN baseline. The authors are also honest about a limitation of their attention design. However, the evaluation protocol has load-bearing weaknesses: the headline numbers are selected from four transformer variants evaluated on the same test set, the baseline Gurobi solutions appear not to be certified optimal, and the deterministic-to-stochastic transfer is asserted rather than validated. These issues need to be addressed before the central claims can be accepted.

major comments (3)
  1. [Section V-B, Table II] This is the central evaluation issue and affects the main quantitative claims.
  2. [Section IV-A, Section V-B] This assumption underpins the entire supervised-learning pipeline and is currently unverified.
  3. [Equation (22), Tables II and III] This affects the interpretation of both the quality and the runtime comparison.
minor comments (5)
  1. [Section IV-B] The motivation for using a transformer mentions its ability 'to capture long term dependencies in sequential data', but the implementation repurposes the sequence dimension to be the feature axis and the feature dimension to be time; attention therefore operates across features rather than time. The authors do acknowledge this limitation, but the framing should be adjusted to avoid overstating the model's temporal modelling capability, or a temporal-attention variant should be considered.
  2. [Section IV-D] The description of the threshold-based post-processing is underspecified: it is not clear whether the mean prediction probability thresholds are computed per class, per instance, or globally over the dataset, nor how the thresholds interact with the class imbalance in a multi-label setting. Please clarify the exact threshold computation and selection rule.
  3. [Table I] The notation for the test set, 'Et ∈ [20, 100] / E5', is difficult to parse. Please explicitly state which EV counts are in the test set (e.g., all integers in [20,100] that are not multiples of 5) and report the number of test instances used in Tables II and III.
  4. [Section V-B] There is a typo in the definition of the multipliers: '{5, 10, 5, 20}' should presumably be '{5, 10, 15, 20}'. Please correct this.
  5. [Abstract and Section I] There are several language issues, including 'combinatorial challenging' in the abstract and 'we proposed' (past tense) in the abstract and introduction. These should be corrected in a final language pass.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the transformer-assisted Gurobi speedup is an externally measured benchmark, not an input-output tautology.

full rationale

The claimed result is an empirical speedup: a transformer is trained on Gurobi-generated optimal binary labels (Section IV-A: 'We then use Gurobi to solve these problems to obtain the optimal binary solutions'), and the assisted solver's runtime and objective are then measured against an independently solved Gurobi baseline using Eqs. (21)-(23). The runtime reduction and quality loss are computed from solver logs and objective values, not from the training labels, so the headline numbers are not equivalent to the model's inputs by construction. The deterministic-to-stochastic labelling shortcut (Section IV-A) and the test-set-based selection of TF15 (Section V-B: 'T F15 will be the best option to assist Gurobi') are evaluation-protocol concerns, not circular reductions; the objective comparison is still independently recalculated by Gurobi. Self-citations to [26] supply formulation details, dataset-construction procedures, and a CNN padding baseline, but the central claim—variable-size transformer pruning—is implemented and tested in this paper, so those citations are not load-bearing. No equation equates a predicted quantity with a measured quantity by definition. Hence no significant circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The empirical claim rests on learned weights, hand-picked thresholds, and assumptions about label quality, distribution transfer, and the power-flow model. These are reasonable for an engineering paper but should be counted when judging the claim's scope and generalizability.

free parameters (2)
  • Post-processing thresholds for class 0 and class 1 predictions = Not stated; computed as mean prediction probability per class on the training data
    Used in Section IV-D to decide which binary variables are fixed before solving. The choice changes feasibility (TF5: 92%, TF15: 100%) and quality, and no principled selection method or sensitivity analysis is provided.
  • Transformer architecture hyperparameters = Not fully reported (2 encoder layers; hidden sizes, heads, learning rate, epochs, batch size missing)
    The architecture is introduced in Sections IV-B and IV-C, but critical training hyperparameters are omitted, so the reported 98% runtime reduction depends on an unspecified tuning process.
assumptions (4)
  • domain assumption Gurobi returns globally optimal binary solutions for all deterministic training instances used as labels.
    Section IV-A builds the training set from Gurobi solutions; if any solve terminates suboptimally, the transformer learns to imitate suboptimal decisions.
  • ad hoc to paper Binary decision patterns learned on deterministic instances transfer to the full stochastic problem.
    Training labels come from a deterministic version of the EVJRS while the test set is the full stochastic version (Section IV-A, Table I). This distribution-shift premise is not validated on any held-out deterministic/stochastic pair.
  • domain assumption LinDistFlow accurately represents line power flows and bus voltage limits for the distribution network.
    The DN operational constraints rely on LinDistFlow (Section III-A). It is a standard linear approximation, but its error profile is not checked against an exact power flow.
  • ad hoc to paper Attention along the feature axis, rather than the time axis, captures the dependencies needed to predict binary decisions.
    The paper repurposes the transformer sequence axis for features to allow variable EV counts (Section IV-B) and explicitly notes this may weaken temporal dependency capture; the empirical adequacy is untested outside the one dataset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transformer-based Deep Learning Model for Joint Routing and Scheduling with Varying Electric Vehicle Numbers." pith.science (2026). https://pith.science/paper/NK7RZOVY

@misc{pith2026250715385,
  author       = {Pith},
  title        = {Pith review of: Transformer-based Deep Learning Model for Joint Routing and Scheduling with Varying Electric Vehicle Numbers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NK7RZOVY}},
  note         = {Machine review of arXiv:2507.15385}
}
read the original abstract

The growing integration of renewable energy sources in modern power systems has introduced significant operational challenges due to their intermittent and uncertain outputs. In recent years, mobile energy storage systems (ESSs) have emerged as a popular flexible resource for mitigating these challenges. Compared to stationary ESSs, mobile ESSs offer additional spatial flexibility, enabling cost-effective energy delivery through the transportation network. However, the widespread deployment of mobile ESSs is often hindered by the high investment cost, which has motivated researchers to investigate utilising more readily available alternatives, such as electric vehicles (EVs) as mobile energy storage units instead. Hence, we explore this opportunity with a MIP-based day-ahead electric vehicle joint routing and scheduling problem in this work. However, solving the problem in a practical setting can often be computationally intractable since the existence of binary variables makes it combinatorial challenging. Therefore, we proposed to simplify the problem's solution process for a MIP solver by pruning the solution search space with a transformer-based deep learning (DL) model. This is done by training the model to rapidly predict the optimal binary solutions. In addition, unlike many existing DL approaches that assume fixed problem structures, the proposed model is designed to accommodate problems with EV fleets of any sizes. This flexibility is essential since frequent re-training can introduce significant computational overhead. We evaluated the approach with simulations on the IEEE 33-bus system coupled with the Nguyen-Dupuis transportation network.

Figures

Figures reproduced from arXiv: 2507.15385 by the authors.

Figure 1
Figure 1. Standard box plots of solution times for 80 random samples of [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Training and inference framework of the proposed DL solution methodology. An overview of the problem model and the proposed transformer-based [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The original TSN (left) proposed in [28] and the modified TSN (right). [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: FFN mechanism for variable input dimensions. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 2
Figure 2. Figure 2: In the feature extraction module, an embedding layer [PITH_FULL_IMAGE:figures/full_fig_p005_2.png]
Figure 5
Figure 5. Figure 5: (a) IEEE 33-bus and (b) Nguyen-Dupuis transportation network. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

31 extracted references · 27 canonical work pages

  1. [1]

    Routing and scheduling of electric buses for resilient restoration of distribution system,

    B. Li, Y . Chen, W. Wei, S. Huang, Y . Xiong, S. Mei, and Y . Hou, “Routing and scheduling of electric buses for resilient restoration of distribution system,” IEEE Transactions on Transportation Electrifica- tion, vol. 7, no. 4, pp. 2414–2428, 2021

  2. [2]

    Global ev outlook 2021,

    I. E. Agency, “Global ev outlook 2021,” https://www.iea.org/reports/global-ev-outlook-2021, 2021

  3. [3]

    Beyond the commute: Unlocking the potential of electric vehicles as future energy storage solutions (vision paper),

    M. A. Cheema, H. Wang, W. Wang, A. N. Toosi, E. Tanin, J. Qi, and H. Samet, “Beyond the commute: Unlocking the potential of electric vehicles as future energy storage solutions (vision paper),” in Proceedings of the 32nd ACM International Conference on Advances in Geographic Information Systems , ser. SIGSPATIAL ’24. New York, NY , USA: Association for Co...

  4. [4]

    Unit commitment considering multiple charging and discharging scenarios of plug-in electric vehicles,

    Z. Yang, K. Li, Q. Niu, and A. Foley, “Unit commitment considering multiple charging and discharging scenarios of plug-in electric vehicles,” in 2015 International Joint Conference on Neural Networks (IJCNN) . IEEE, 2015, pp. 1–8

  5. [5]

    A binary symmetric based hybrid meta-heuristic method for solving mixed integer unit commitment problem integrating with significant plug-in electric vehicles,

    Z. Yang, K. Li, Y . Guo, S. Feng, Q. Niu, Y . Xue, and A. Foley, “A binary symmetric based hybrid meta-heuristic method for solving mixed integer unit commitment problem integrating with significant plug-in electric vehicles,” Energy, vol. 170, pp. 889–905, 2019

  6. [6]

    Ev scheduling framework for peak demand manage- ment in lv residential networks,

    S. Rafique, M. S. H. Nizami, U. B. Irshad, M. J. Hossain, and S. C. Mukhopadhyay, “Ev scheduling framework for peak demand manage- ment in lv residential networks,” IEEE Systems Journal , vol. 16, no. 1, pp. 1520–1528, 2021

  7. [7]

    A two-stage multi-objective stochastic optimization strategy to minimize cost for electric bus depot operators,

    S. Rafique, M. Nizami, U. Irshad, M. Hossain, and S. Mukhopadhyay, “A two-stage multi-objective stochastic optimization strategy to minimize cost for electric bus depot operators,” Journal of Cleaner Production , vol. 332, p. 129856, 2022

  8. [8]

    Trilevel mixed integer opti- mization for day-ahead spinning reserve management of electric vehicle aggregator with uncertainty,

    W. Liu, S. Chen, Y . Hou, and Z. Yang, “Trilevel mixed integer opti- mization for day-ahead spinning reserve management of electric vehicle aggregator with uncertainty,” IEEE Transactions on Smart Grid, vol. 13, no. 1, pp. 613–625, 2022

Show all 31 references
  1. [9]

    Joint routing and scheduling for electric vehicles in smart grids with v2g,

    A. Trivi ˜no-Cabrera, J. A. Aguado, and S. de la Torre, “Joint routing and scheduling for electric vehicles in smart grids with v2g,” Energy, vol. 175, pp. 113–122, 2019. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0360544219303901

  2. [10]

    Joint routing and charging problem of multiple electric vehicles: A fast optimization algorithm,

    C. Yao, S. Chen, and Z. Yang, “Joint routing and charging problem of multiple electric vehicles: A fast optimization algorithm,” IEEE Transactions on Intelligent Transportation Systems , vol. 23, no. 7, pp. 8184–8193, 2021

  3. [11]

    Congestion-aware dynamic optimal traffic power flow in coupled transportation power systems,

    T. Zhao, H. Yan, X. Liu, and Z. Ding, “Congestion-aware dynamic optimal traffic power flow in coupled transportation power systems,” IEEE Transactions on Industrial Informatics , vol. 19, no. 2, pp. 1833– 1843, 2022

  4. [12]

    Collaborative ev routing and charging scheduling with power distribution and traffic networks interaction,

    J. Liu, G. Lin, S. Huang, Y . Zhou, C. Rehtanz, and Y . Li, “Collaborative ev routing and charging scheduling with power distribution and traffic networks interaction,” IEEE Transactions on Power Systems , vol. 37, no. 5, pp. 3923–3936, 2022

  5. [13]

    Optimal routing and power management of electric vehicles in coupled power distribution and transportation systems,

    S. Aghajan-Eshkevari, M. T. Ameli, and S. Azad, “Optimal routing and power management of electric vehicles in coupled power distribution and transportation systems,” Applied Energy, vol. 341, p. 121126, 2023

  6. [14]

    Equilibrium analysis of electricity markets with day-ahead market power mitigation and real-time intercept bidding,

    R. K. Bansal, Y . Chen, P. You, and E. Mallada, “Equilibrium analysis of electricity markets with day-ahead market power mitigation and real-time intercept bidding,” in Proceedings of the Thirteenth ACM International Conference on Future Energy Systems , 2022, pp. 47–62

  7. [15]

    Gurobi Optimizer Reference Manual,

    Gurobi Optimization, LLC, “Gurobi Optimizer Reference Manual,”

  8. [16]

    An efficient method for computing traffic equilibria in networks with asymmetric transportation costs,

    S. Nguyen and C. Dupuis, “An efficient method for computing traffic equilibria in networks with asymmetric transportation costs,” Trans- portation Science, vol. 18, no. 2, pp. 185–202, 1984

  9. [17]

    Q-learning-based model predictive control for energy management in residential aggregator,

    K. Ojand and H. Dagdougui, “Q-learning-based model predictive control for energy management in residential aggregator,” IEEE Transactions on Automation Science and Engineering , vol. 19, no. 1, pp. 70–81, 2021

  10. [18]

    Siphyr: An end-to-end learning-based optimization framework for dynamic grid reconfiguration,

    R. Haider, A. Annaswamy, B. Dey, and A. Chakraborty, “Siphyr: An end-to-end learning-based optimization framework for dynamic grid reconfiguration,” IEEE Transactions on Smart Grid , 2024

  11. [19]

    Optimal control of microgrids with multi-stage mixed-integer nonlinear programming guided q-learning algorithm,

    Y . Yoldas, S. Goren, and A. Onen, “Optimal control of microgrids with multi-stage mixed-integer nonlinear programming guided q-learning algorithm,” Journal of Modern Power Systems and Clean Energy, vol. 8, no. 6, pp. 1151–1159, 2020

  12. [20]

    A hybrid approach for home energy management with imitation learning and online optimization,

    S. Gao, R. bin Zulkifli Lee, Z. Huang, C. Xiang, M. Yu, K. T. Tan, and T. H. Lee, “A hybrid approach for home energy management with imitation learning and online optimization,” IEEE Transactions on Industrial Informatics, vol. 20, no. 3, pp. 4527–4539, 2023

  13. [21]

    Combining deep learning and optimization for preventive security-constrained dc optimal power flow,

    A. Velloso and P. Van Hentenryck, “Combining deep learning and optimization for preventive security-constrained dc optimal power flow,” IEEE Transactions on Power Systems , vol. 36, no. 4, pp. 3618–3628, 2021

  14. [22]

    Machine learning-additional decision constraints for improved milp day-ahead unit commitment method,

    M. I. A. Shekeew and B. Venkatesh, “Machine learning-additional decision constraints for improved milp day-ahead unit commitment method,” IEEE Access, vol. 11, pp. 111 976–111 990, 2023

  15. [23]

    Reinforcement learning and mixed-integer programming for power plant scheduling in low carbon systems: Comparison and hybridisation,

    C. O’Malley, P. de Mars, L. Badesa, and G. Strbac, “Reinforcement learning and mixed-integer programming for power plant scheduling in low carbon systems: Comparison and hybridisation,” Applied Energy , vol. 349, p. 121659, 2023

  16. [24]

    Learning-assisted variables reduc- tion method for large-scale milp unit commitment,

    M. I. A. Shekeew and B. Venkatesh, “Learning-assisted variables reduc- tion method for large-scale milp unit commitment,” IEEE Open Access Journal of Power and Energy , vol. 10, pp. 245–258, 2023

  17. [25]

    Data- augmentation acceleration framework by graph neural network for near- optimal unit commitment,

    L. Wei, X. Ai, J. Fang, S. Cui, L. Gao, K. Li, and J. Wen, “Data- augmentation acceleration framework by graph neural network for near- optimal unit commitment,” Applied Energy, vol. 377, p. 124332, 2025

  18. [26]

    Joint optimisation of electric vehicle routing and scheduling: A deep learning-driven approach for dynamic fleet sizes,

    J. K. Yap, V . M. Baskaran, W. S. Tan, Z. Y . Ding, H. Wang, and D. L. Dowe, “Joint optimisation of electric vehicle routing and scheduling: A deep learning-driven approach for dynamic fleet sizes,” in 2025 International joint conference on neural networks (IJCNN) . IEEE, 2025...

  19. [27]

    An interval power flow method based on linearized distflow equations for radial distribution systems,

    J. Du, J. Tian, Z. Wu, A. Li, G. Abbas, and Q. Sun, “An interval power flow method based on linearized distflow equations for radial distribution systems,” in 2020 12th IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC) . IEEE, 2020, pp. 1–5

  20. [28]

    Battery-based energy stor- age transportation for enhancing power system economics and security,

    Y . Sun, Z. Li, M. Shahidehpour, and B. Ai, “Battery-based energy stor- age transportation for enhancing power system economics and security,” IEEE Transactions on Smart Grid , vol. 6, no. 5, pp. 2395–2402, 2015

  21. [29]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems , vol. 30, 2017

  22. [30]

    Asymmetric loss for multi-label classification,

    T. Ridnik, E. Ben-Baruch, N. Zamir, A. Noy, I. Friedman, M. Protter, and L. Zelnik-Manor, “Asymmetric loss for multi-label classification,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 82–91

  23. [2023]

    Available: https://www.gurobi.com

    [Online]. Available: https://www.gurobi.com

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.