Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learning

Amber Cassimon; Kevin Mets; Siegfried Mercelis

REVIEW 5 major objections 5 minor 35 references

Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learning

T0 review · 5 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 Pretraining a neural architecture search agent on one vision task gives it a better start on another task in almost every case tested, cutting the training needed.

desk verdict A useful empirical transfer study whose significance claims rest on impossible confidence intervals; the qualitative direction is plausible, but the statistics need to be redone before the strong claims can be trusted. read the letter →

arxiv 2412.01420 v2 pith:RUG5HWUQ submitted 2024-12-02 cs.LG

Amber Cassimon , Siegfried Mercelis , Kevin Mets This is my paper · ORCID

classification cs.LG

keywords neuralarchitecturesearchreinforcementlearningtransferfine-tuningzero-shotTransNASBench-101rewardshapingApe-X

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

Neural architecture search (NAS) is computationally expensive because each new task normally requires training the search agent from scratch. This paper tests whether a reinforcement-learning NAS agent that improves architectures incrementally can instead be pretrained on one computer-vision task and then transferred to another. Using four tasks from the TransNASBench-101 benchmark, the authors report that transferred agents match or beat from-scratch agents in final performance for three of the four target tasks, and that pretrained agents reach a given performance level in substantially fewer time steps. If the result holds, the upfront cost of training NAS agents could be amortized across many tasks, which matters because such agents can take dozens to hundreds of GPU-hours to train.

What carries the argument

The central object is the incremental-improvement RL NAS agent from previous work: a transformer-encoded policy, trained with Ape-X, double Q-learning, dueling heads, three-step bootstrapping with partial episode bootstrapping, and prioritized experience replay, that repeatedly proposes small edits to the current architecture until the episode ends. The transfer mechanism is parameter initialization: the entire learned policy, including the architecture encoder and transformer layers, is copied from the source task and then trained on the target task. The paper's one new component is a gamma reward transform, $R'(s,a)=R(s,a)^{\gamma}$, applied to the semantic-segmentation task (with $\gamma=0.478$ chosen to maximize reward spread), because the raw mIoU reward is concentrated near zero and otherwise gives the agent too little learning signal.

What would settle it

Recompute the 95% bootstrap confidence intervals from the raw per-run scores for each cell; if, as the reported numbers suggest, three runs with a standard deviation of about 0.038 yield an interval only about 0.001 wide, the bootstrap procedure is not producing valid intervals, and the non-overlap significance claims fail. A complementary check would be to repeat the 4x4 transfer matrix with ten or more seeds and test whether transferred agents still consistently beat from-scratch agents on three of the four target tasks.

Watch

Extended reading notes

Core claim

The paper's central claim is that a single RL-based NAS policy, pretrained on one task, is a better initialization for another task than random initialization, and that this holds across different source-target pairs despite the tasks having low ranking correlation with each other. Under both a fine-tuning regime ($10^6$ target steps) and a full retraining regime ($10^7$ target steps), the transferred agents achieve validation performance that meets or beats the from-scratch baseline on three of the four target tasks, with non-overlapping bootstrapped 95% confidence intervals cited as evidence of statistical significance. The same data show that the crossover point at which a transferred agent surpasses a from-scratch agent's $10^7$-step performance often occurs orders of magnitude earlier in training. The paper attributes part of the benefit to the shared search space: all tasks use the same architecture encoding, so the early embedding layers can transfer largely unchanged, while later layers are retrained.

Load-bearing premise

The load-bearing premise is that the bootstrapped 95% confidence intervals, computed from only three random-seed runs, are valid; if the interval calculation is wrong, the non-overlap argument used to claim statistical significance collapses, even though the point estimates may still favor transferred agents.

Editorial extensions

If this is right

Pretraining an RL-based NAS agent on one task can shorten target-task training by orders of magnitude: in several source-target pairs the transferred agent matches the from-scratch agent's $10^7$-step performance within $10^3$ to $10^5$ steps.
Transfer benefits appear even between tasks with low ranking correlation, so the shared search space alone may explain part of the gain rather than task similarity.
Zero-shot transfer is consistently weaker, while even $10^6$ fine-tuning steps recover most of the benefit, so some target-task interaction is still required.
The same training-cost observation now holds on three standardized NAS benchmarks, supporting the view that search-space size is not the main driver of the agent's convergence time.
If the pretrained initialization is understood as a better starting point rather than a warm start, the effective cost of applying NAS to a new task may drop substantially, making RL-based NAS practical in settings where full training is unaffordable.

Reading between the lines

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

The reported numbers hint that the bootstrapped confidence intervals may be too narrow for three runs (for example, a standard deviation of 0.038 with N=3 should not yield an interval width near 0.001); recomputing them with a correct percentile bootstrap could either strengthen or weaken the significance conclusions.
A natural testable extension would be freezing the architecture-embedding layers after pretraining and retraining only the policy head, which the paper's shared-search-space argument suggests should retain most of the transfer benefit at lower cost.
Because only Ape-X was tested, the authors' conclusions are tied to that algorithm's exploration behavior; trying a sample-efficient method would show whether the transfer advantage is a property of the task or of the optimizer.
The paper's two explanations for the effect—from-scratch runs getting stuck in local optima and a grokking-like late improvement—make distinguishable predictions about when transferred curves overtake scratch curves, so tracking per-run trajectories could separate them.

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.

Desk Editor's Note

read the letter

The genuinely new thing here is the first evaluation of fine-tuning transfer for the incremental RL NAS agent from Cassimon et al. (2024). The setup is clean: four tasks from Trans-NASBench-101, three transfer regimes, and they measure both final performance and time-to-match a from-scratch baseline. They also include a sensible limitations section and are honest about the sample-efficiency weakness of Ape-X. On the empirical side, the point estimates consistently favor transferred agents, and the cross-over time results, if the numbers are right, could be practically useful for someone deciding whether to reuse a pretrained agent.

The soft spot is not small. The reported 95% confidence intervals in Figures 5–7 are arithmetically impossible. For example, the autoencoder diagonal in Figure 5 reports mean 0.518 with standard deviation 0.038 across N=3 runs, but a confidence interval of [0.518, 0.519], width 0.001. With three runs, the standard error alone is about 0.022, and no bootstrap procedure can produce a percentile interval narrower than the spread of the three sample means. The same pattern appears throughout those figures. Since the paper explicitly uses CI non-overlap as its significance criterion, all of the “statistically significant” claims about transfer benefits and shortened training are unsupported as reported. The cross-over time intervals in Figures 12–13 are less obviously impossible, but they also look suspiciously narrow given the accompanying standard deviations.

The point estimates still support the qualitative direction: transferred agents usually match or beat from-scratch agents at the final evaluation, and they often reach the from-scratch performance level faster. So I would not dismiss the paper’s main hypothesis. But the strength of the central claim, as stated in the abstract, is not backed by the statistics. The gamma reward shaping is a fitted parameter with a single value chosen on the validation set, which is a minor concern compared to the CI problem.

For a reader in the NAS subfield, this is a worthwhile empirical question, and the paper is a legitimate extension of prior work rather than a breakthrough. It deserves a serious referee, but not a quick acceptance. The authors should recompute all of the intervals using a correct bootstrap or a proper small-sample test, report effect sizes instead of relying on overlap, and ideally release code and raw data so the numbers can be checked. I would not cite the quantitative significance claims until that is fixed, but the paper is worth engaging with in review rather than desk rejecting.

Referee Report

5 major / 5 minor

Summary. The paper studies whether a reinforcement-learning-based neural architecture search (NAS) agent, originally introduced by Cassimon et al., can be reused across tasks via simple transfer learning. Four TransNASBench-101 tasks are selected (autoencoder, class-object, room-layout, segmentsemantic), and three regimes are compared: zero-shot transfer, fine-tuning for 1e6 timesteps, and retraining for 1e7 timesteps. The central empirical claims are that pretraining on one task benefits final performance on another task in all but one target task, and that transferred agents can be trained to match a from-scratch agent in significantly fewer timesteps. The paper uses point estimates, standard deviations, and 95% bootstrap confidence intervals from N=3 random seeds, and uses confidence-interval non-overlap as its statistical significance criterion.

Significance. If the qualitative findings hold, the paper addresses a practical bottleneck of RL-based NAS: the high cost of training a new agent for every task. The study is clearly scoped, uses a standardized benchmark, and documents the agent architecture and hyperparameters in enough detail to be reproducible in principle. It also offers two plausible mechanistic hypotheses (local optima and a grokking-like effect) for why transfer helps. The paper's usefulness, however, depends on the reliability of its statistical significance claims, and those claims are currently undermined by internally inconsistent confidence intervals. With corrected statistics and appropriately softened language, the direction of the results would still be informative; the study would then be a useful empirical data point on cross-task transfer for RL-based NAS agents.

major comments (5)

[§4.1, Figures 5–7] The reported 95% bootstrap confidence intervals are arithmetically inconsistent with the stated sample size and standard deviations, and the non-overlap argument built on them is therefore unsupported. For example, in Figure 5 the autoencoder from-scratch cell reports mean=0.518, SD=0.038, and CI=[0.518,0.519], a width of about 0.001. With N=3, the standard error of the mean is 0.038/sqrt(3)≈0.022, and a bootstrap percentile interval over three observations cannot be narrower than the range of the three values. The same pattern appears throughout Figures 5–7 and in Figures 12–13. Because Sections 4.1 and 4.3 explicitly use CI non-overlap as the significance criterion, the claims 'indicating a statistically significant difference' and 'statistically significant improvement' rest on invalid intervals. The authors should recompute intervals with a correct small-sample method, report the method and number of resamples, and revise the significance language accordingly.
[§4.3, Figures 10–13] The cross-over time statistics are not credible as reported and need to be recomputed. In Figure 13, the autoencoder→autoencoder cell gives 5.359×10^6 ± 6.801×10^2 timesteps with CI [4.820×10^6, 6.041×10^6]; the CI width is about 1.2×10^6, which is three orders of magnitude larger than the reported standard deviation, and the SD value appears to be copied from the wall-clock version in Figure 12. Moreover, the 'all-to-all comparison' creates 9 pairs from 3×3 runs, but these pairs are not independent because the same runs are reused; a bootstrap over pairs will understate uncertainty relative to a bootstrap over independent seeds. Without corrected estimates, the claim that pretraining 'significantly' shortens training is not supported by the reported evidence.
[§4.1 and §4.2] Even apart from the CI arithmetic, the strength of the language is disproportionate to a sample size of N=3 per condition. The class-object case is acknowledged as non-significant, yet the surrounding text states that transferred agents outperform from-scratch agents in 'all but 1 task' and that the effects occur 'regardless of the source or target task.' Some of the reported differences are also small in absolute terms (e.g., 49.2% vs. 49.5% for class-object in Figure 5). The paper should report effect sizes, present per-seed values or at least the raw data necessary to verify them, and phrase the conclusions in terms of the direction of the point estimates rather than binary statistical significance.
[§4, Figures 8–9] The shaded 95% confidence bands in the training-curve plots are computed from the same N=3 runs, and consecutive timesteps of the same run are highly correlated. The manuscript does not state how the moving-average filter, the multiple timesteps, and the repeated measures are accounted for in the band computation. Without this clarification, the visual separation of the bands should not be interpreted as evidence of statistically significant differences in convergence speed. This is a load-bearing point for the 'shortened training' claim and should either be rigorously justified or de-emphasized.
[§3.2] The gamma reward shaping parameter for the segmentsemantic task was selected empirically by sweeping gamma and choosing the value that maximizes validation reward spread. Because the same task is then used in the transfer comparisons, the paper should clarify whether this selection is a form of model selection on the evaluation task and, if so, how it affects the comparison between from-scratch and transferred agents on that task. At minimum, the sweep range, the exact definition of 'spread,' and the criterion for 'optimal' should be stated explicitly.

minor comments (5)

[Throughout] There are several typos and inconsistencies, including 'trained froms scratch' in the captions of Figures 10, 12, and 13, 'CIF AR-10' in the related-work section, 'up-front of cost' in the introduction, and inconsistent use of 'class-object' versus 'class_object' between text and figures.
[§4.3] The 'all-to-all comparison scheme' is only described in one sentence. Please define precisely how the cross-over point is computed, how runs that never reach the reference performance within the 1e7-timestep budget are handled, and whether censored runs are excluded or assigned a maximum value.
[§5.5] The data availability statement says data are available 'upon reasonable request.' For a study whose conclusions rest on small-sample statistics, depositing the per-seed results and the code used to compute the confidence intervals would substantially increase confidence in the reported numbers.
[§3.1, Figure 1] Figure 1 reports Kendall's tau correlations between tasks, but the text does not say what is being correlated (architectures? validation scores? a sample size?) or whether the values are computed on the validation or test set. Please add this information.
[References] The reference for the Adam optimizer is incomplete ('Kingma D (2014) Adam: a method for stochastic optimization'), and some reference entries have inconsistent formatting. These should be cleaned up before publication.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: transfer conclusions are empirical comparisons against from-scratch baselines; self-citations supply the agent and hyperparameters, not the target result.

full rationale

The paper's central claims are empirical: pretrained agents are compared with from-scratch agents on final performance and on cross-over training time (Sections 4.1 and 4.3, Figures 5 and 8-13). The only introduced equation is the gamma reward reshaping R'(s,a)=R(s,a)^γ (Section 3.2); γ=0.478 is selected by maximizing reward spread for the segmentsemantic task and is applied equally to from-scratch and transferred agents, so it cannot by construction produce the transfer benefit. The Cassimon et al. (2024) citations provide the agent architecture, observation processing, and most hyperparameters ('The hyperparameters we use are mostly the same as those used by Cassimon et al (2024)'), i.e., a tool rather than evidence for the transfer conclusion; the transfer result is established by the experiments reported here. The paper explicitly notes limitations (only Ape-X was tested; only four Trans-NASBench-101 tasks; same search-space assumption), which are generality caveats, not circularity. The reported 95% confidence intervals appear arithmetically inconsistent with N=3 (e.g., Figure 5, autoencoder from-scratch: mean 0.518, std 0.038, CI [0.518,0.519]), which threatens the statistical significance claims built on CI non-overlap, but this is a statistical correctness concern rather than an equivalence-by-construction between inputs and outputs. No fitted parameter is renamed as a prediction, and no equation in the paper reduces the transfer advantage to its own definition.

Assumptions & free parameters 1 free parameters · 2 assumptions · 0 invented entities

The central claim depends on the benchmark's fidelity and on the representativeness of the Cassimon et al agent. One free parameter (gamma) is fitted to validation data for the semantic segmentation task.

free parameters (1)

gamma reward shaping parameter = 0.478
Chosen empirically to maximize the spread of gamma-transformed semantic segmentation validation rewards (Section 3.2, Figure 4). Used for all segment semantic training runs, so it affects the transfer results for that task.

assumptions (2)

domain assumption Trans-NASBench-101 provides faithful performance estimates for architectures on the four selected tasks.
The entire evaluation uses this benchmark as ground truth for rewards and final metrics without independent verification (Section 3.1).
domain assumption The RL-based NAS agent from Cassimon et al (2024) is a representative instance of RL-based NAS agents.
The paper tests only this single agent and acknowledges in Section 5 that conclusions may not hold for other RL algorithms.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learning." pith.science (2026). https://pith.science/paper/RUG5HWUQ

@misc{pith2026241201420,
  author       = {Pith},
  title        = {Pith review of: Task Adaptation of Reinforcement Learning-based NAS Agents through Transfer Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RUG5HWUQ}},
  note         = {Machine review of arXiv:2412.01420}
}

read the original abstract

Recently, a novel paradigm has been proposed for reinforcement learning-based NAS agents, that revolves around the incremental improvement of a given architecture. We assess the abilities of such reinforcement learning agents to transfer between different tasks. We perform our evaluation using the Trans-NASBench-101 benchmark, and consider the efficacy of the transferred agents, as well as how quickly they can be trained. We find that pretraining an agent on one task benefits the performance of the agent in another task in all but 1 task when considering final performance. We also show that the training procedure for an agent can be shortened significantly by pretraining it on another task. Our results indicate that these effects occur regardless of the source or target task, although they are more pronounced for some tasks than for others. Our results show that transfer learning can be an effective tool in mitigating the computational cost of the initial training procedure for reinforcement learning-based NAS agents.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 26 canonical work pages

[1]

In: International Conference on Learning Representations, ://openreview.net/forum?id=S1c2cvqee

Baker B, Gupta O, Naik N, et al (2017) Designing neural network architectures using reinforcement learning. In: International Conference on Learning Representations, ://openreview.net/forum?id=S1c2cvqee

work page 2017
[2]

Neural Computing and Applications doi:10.1007/s00521-024-10445-2, ://doi.org/10.1007/s00521-024-10445-2

Cassimon A, Mercelis S, Mets K (2024) Scalable reinforcement learning-based neural architecture search. Neural Computing and Applications doi:10.1007/s00521-024-10445-2, ://doi.org/10.1007/s00521-024-10445-2

work page doi:10.1007/s00521-024-10445-2 2024
[3]

Devlin J, Chang MW, Lee K, et al (2019) BERT : Pre-training of deep bidirectional transformers for language understanding. In: Burstein J, Doran C, Solorio T (eds) Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). Association fo...

doi:10.18653/v1/n19-1423 2019
[4]

In: International Conference on Learning Representations, ://openreview.net/forum?id=KEQl-MZ5fg7

Ding M, Huo Y, Lu H, et al (2022) Learning versatile neural architectures by propagating network codes. In: International Conference on Learning Representations, ://openreview.net/forum?id=KEQl-MZ5fg7

work page 2022
[5]

://arxiv.org/abs/2001.00326

Dong X, Yang Y (2020) NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search pp 1--16. ://arxiv.org/abs/2001.00326

arXiv 2020
[6]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 5251--5260

Duan Y, Chen X, Xu H, et al (2021) Transnas-bench-101: Improving transferability and generalizability of cross-task neural architecture search. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 5251--5260

work page 2021
[7]

In: International Conference on Learning Representations, ://openreview.net/forum?id=ByME42AqK7

Elsken T, Metzen JH, Hutter F (2019) Efficient multi-objective neural architecture search via lamarckian evolution. In: International Conference on Learning Representations, ://openreview.net/forum?id=ByME42AqK7

work page 2019
[8]

In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

He K, Zhang X, Ren S, et al (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

work page 2016

Show all 35 references

[9]

In: The Twelfth International Conference on Learning Representations, ://openreview.net/forum?id=qPloNoDJZn

He Z, Shu Y, Dai Z, et al (2024) Robustifying and boosting training-free neural architecture search. In: The Twelfth International Conference on Learning Representations, ://openreview.net/forum?id=qPloNoDJZn

2024
[10]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 11881--11891

Huang M, Huang Z, Li C, et al (2022) Arch-graph: Acyclic architecture relation predictor for task-transferable neural architecture search. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 11881--11891

2022
[11]

://arxiv.org/abs/1608.08614, https://arxiv.org/abs/1608.08614 1608.08614

Huh M, Agrawal P, Efros AA (2016) What makes imagenet good for transfer learning? CoRR abs/1608.08614. ://arxiv.org/abs/1608.08614, https://arxiv.org/abs/1608.08614 1608.08614

2016 arXiv
[12]

In: Kober J, Ramos F, Tomlin C (eds) Proceedings of the 2020 Conference on Robot Learning, Proceedings of Machine Learning Research, vol 155

Julian R, Swanson B, Sukhatme G, et al (2021) Never stop learning: The effectiveness of fine-tuning in robotic reinforcement learning. In: Kober J, Ramos F, Tomlin C (eds) Proceedings of the 2020 Conference on Robot Learning, Proceedings of Machine Learning Research, vol 155. ...

2021
[13]

Nature 596(7873):583--589

Jumper J, Evans R, Pritzel A, et al (2021) Highly accurate protein structure prediction with alphafold. Nature 596(7873):583--589

2021
[14]

Khosla A, Jayadevaprakash N, Yao B, et al (2011) Novel dataset for fine-grained image categorization. In: First Workshop on Fine-Grained Visual Categorization, IEEE Conference on Computer Vision and Pattern Recognition, Colorado Springs, CO, ://vision.stanford.edu/aditya86/Ima...

2011
[15]

arXiv preprint arXiv:14126980

Kingma D (2014) Adam: a method for stochastic optimization. arXiv preprint arXiv:14126980

2014
[16]

Krizhevsky A, Hinton G, et al (2009) Learning multiple layers of features from tiny images

2009
[17]

Transactions on Machine Learning Research ://openreview.net/forum?id=ok18jj7cam

Li M, Liu JY, Sigal L, et al (2023) Graph PNAS : Learning probabilistic graph generators for neural architecture search. Transactions on Machine Learning Research ://openreview.net/forum?id=ok18jj7cam

2023
[18]

In: International Conference on Learning Representations, ://openreview.net/forum?id=S1eYHoC5FX

Liu H, Simonyan K, Yang Y (2019) DARTS : Differentiable architecture search. In: International Conference on Learning Representations, ://openreview.net/forum?id=S1eYHoC5FX

2019
[19]

Proceedings of the AAAI Conference on Artificial Intelligence 37(8):9180--9189

Mills KG, Niu D, Salameh M, et al (2023) Aio-p: Expanding neural performance predictors beyond image classification. Proceedings of the AAAI Conference on Artificial Intelligence 37(8):9180--9189. doi:10.1609/aaai.v37i8.26101, ://ojs.aaai.org/index.php/AAAI/article/view/26101

2023 doi
[20]

In: Dy J, Krause A (eds) Proceedings of the 35th International Conference on Machine Learning, Proceedings of Machine Learning Research, vol 80

Pardo F, Tavakoli A, Levdik V, et al (2018) Time limits in reinforcement learning. In: Dy J, Krause A (eds) Proceedings of the 35th International Conference on Machine Learning, Proceedings of Machine Learning Research, vol 80. PMLR, pp 4045--4054, ://proceedings.mlr.press/v80...

2018
[21]

In: Proceedings of the 35th International Conference on Machine Learning, pp 4095--4104, ://proceedings.mlr.press/v80/pham18a.html

Pham H, Guan M, Zoph B, et al (2018) Efficient neural architecture search via parameters sharing. In: Proceedings of the 35th International Conference on Machine Learning, pp 4095--4104, ://proceedings.mlr.press/v80/pham18a.html

2018
[22]

Power A, Burda Y, Edwards H, et al (2021) Grokking: Generalization beyond overfit-ting on small algorithmic datasets. In: 1st Mathematical Reasoning in General Artificial Intelligence Workshop at the International Conference on Learning Representations, ://mathai-iclr.github.i...

2021
[23]

Mechanical Systems and Signal Processing 186:109884

Qian Q, Qin Y, Luo J, et al (2023) Deep discriminative transfer learning network for cross-machine fault diagnosis. Mechanical Systems and Signal Processing 186:109884. doi:https://doi.org/10.1016/j.ymssp.2022.109884, ://www.sciencedirect.com/science/article/pii/S0888327022009529

2023
[24]

In: Proceedings of the 4th Inter national Conference on Learning Representations, ICLR

Schaul T, Quan J, Antonoglou I, et al (2016) Prioritized experience replay [c/ol]. In: Proceedings of the 4th Inter national Conference on Learning Representations, ICLR

2016
[25]

In: International conference on machine learning, Pmlr, pp 6105--6114

Tan M, Le Q (2019) Efficientnet: Rethinking model scaling for convolutional neural networks. In: International conference on machine learning, Pmlr, pp 6105--6114

2019
[26]

Tan M, Chen B, Pang R, et al (2019) Mnasnet: Platform-aware neural architecture search for mobile. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), ://openaccess.thecvf.com/content\_CVPR\_2019/html/Tan\_MnasNet\_Platform-Aware\_Neur...

2019
[27]

In: Koyejo S, Mohamed S, Agarwal A, et al (eds) Advances in Neural Information Processing Systems, vol 35

Tu R, Roberts N, Khodak M, et al (2022) Nas-bench-360: Benchmarking neural architecture search on diverse tasks. In: Koyejo S, Mohamed S, Agarwal A, et al (eds) Advances in Neural Information Processing Systems, vol 35. Curran Associates, Inc., pp 12380--12394, ://proceedings....

2022
[28]

In: Guyon I, Luxburg UV, Bengio S, et al (eds) Advances in Neural Information Processing Systems, vol 30

Vaswani A, Shazeer N, Parmar N, et al (2017) Attention is all you need. In: Guyon I, Luxburg UV, Bengio S, et al (eds) Advances in Neural Information Processing Systems, vol 30. Curran Associates, Inc., ://proceedings.neurips.cc/paper\_files/paper/2017/file/3f5ee243547dee91fbd...

2017
[29]

Machine learning 8:229--256

Williams RJ (1992) Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine learning 8:229--256

1992
[30]

In: Chaudhuri K, Salakhutdinov R (eds) Proceedings of the 36th International Conference on Machine Learning, Proceedings of Machine Learning Research, vol 97

Ying C, Klein A, Christiansen E, et al (2019) NAS -bench-101: Towards reproducible neural architecture search. In: Chaudhuri K, Salakhutdinov R (eds) Proceedings of the 36th International Conference on Machine Learning, Proceedings of Machine Learning Research, vol 97. Pmlr, L...

2019
[31]

IEEE Transactions on Evolutionary Computation 28(3):682--695

Zhou X, Wang Z, Feng L, et al (2024) Toward evolutionary multitask convolutional neural architecture search. IEEE Transactions on Evolutionary Computation 28(3):682--695. doi:10.1109/tevc.2023.3348475

2024
[32]

IEEE Transactions on Pattern Analysis and Machine Intelligence 45(11):13344--13362

Zhu Z, Lin K, Jain AK, et al (2023) Transfer learning in deep reinforcement learning: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence 45(11):13344--13362. doi:10.1109/TPAMI.2023.3292075

2023
[33]

In: International Conference on Learning Representations, ://openreview.net/forum?id=r1Ue8Hcxg

Zoph B, Le Q (2017) Neural architecture search with reinforcement learning. In: International Conference on Learning Representations, ://openreview.net/forum?id=r1Ue8Hcxg

2017
[34]

, " * write output.state after.block = add.period write newline

ENTRY address archive author booktitle chapter doi edition editor eid eprint howpublished institution journal key keywords month note number organization pages publisher school series title type url volume year archivePrefix primaryClass adsurl adsnote version label extra.labe...
[35]

write newline

" write newline "" before.all 'output.state := FUNCTION add.period duplicate empty 'skip "." * add.blank if FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap dupl...

Pith tools

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

[1] [1]

In: International Conference on Learning Representations, ://openreview.net/forum?id=S1c2cvqee

Baker B, Gupta O, Naik N, et al (2017) Designing neural network architectures using reinforcement learning. In: International Conference on Learning Representations, ://openreview.net/forum?id=S1c2cvqee

work page 2017

[2] [2]

Neural Computing and Applications doi:10.1007/s00521-024-10445-2, ://doi.org/10.1007/s00521-024-10445-2

Cassimon A, Mercelis S, Mets K (2024) Scalable reinforcement learning-based neural architecture search. Neural Computing and Applications doi:10.1007/s00521-024-10445-2, ://doi.org/10.1007/s00521-024-10445-2

work page doi:10.1007/s00521-024-10445-2 2024

[3] [3]

Devlin J, Chang MW, Lee K, et al (2019) BERT : Pre-training of deep bidirectional transformers for language understanding. In: Burstein J, Doran C, Solorio T (eds) Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). Association fo...

doi:10.18653/v1/n19-1423 2019

[4] [4]

In: International Conference on Learning Representations, ://openreview.net/forum?id=KEQl-MZ5fg7

Ding M, Huo Y, Lu H, et al (2022) Learning versatile neural architectures by propagating network codes. In: International Conference on Learning Representations, ://openreview.net/forum?id=KEQl-MZ5fg7

work page 2022

[5] [5]

://arxiv.org/abs/2001.00326

Dong X, Yang Y (2020) NAS-Bench-201: Extending the Scope of Reproducible Neural Architecture Search pp 1--16. ://arxiv.org/abs/2001.00326

arXiv 2020

[6] [6]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 5251--5260

Duan Y, Chen X, Xu H, et al (2021) Transnas-bench-101: Improving transferability and generalizability of cross-task neural architecture search. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp 5251--5260

work page 2021

[7] [7]

In: International Conference on Learning Representations, ://openreview.net/forum?id=ByME42AqK7

Elsken T, Metzen JH, Hutter F (2019) Efficient multi-objective neural architecture search via lamarckian evolution. In: International Conference on Learning Representations, ://openreview.net/forum?id=ByME42AqK7

work page 2019

[8] [8]

In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

He K, Zhang X, Ren S, et al (2016) Deep residual learning for image recognition. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

work page 2016