Elastic Machine Learning Algorithms in Amazon SageMaker

Liberty, Edo; Karnin, Zohar; Xiang, Bing; Rouesnel, Laurence; Coskun, Baris; Nallapati, Ramesh; Delgado, Julio; Sadoughi, Amir; Astashonok, Yury; Das, Piali; Balioglu, Can; Chakravarty, Saswata; Jha, Madhav; Gautier, Philip; Arpin, D; Januschowski, Tim; Flunkert, Valentín; Wang, Yuyang; Gasthaus, Jan; Stella, Lorenzo; Rangapuram, Syama Sundar; Salinas, David; Schelter, Sebastian; Smola, Alex

doi:10.1145/3318464.3386126

Cited by 85 publications

(44 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…We implement the model in the open-source GluonTS 4 [1] library. The experiments are done using AWS Sage-Maker [17].…”

Section: Methodsmentioning

confidence: 99%

Learning Quantile Functions without Quantile Crossing for Distribution-free Time Series Forecasting

Park¹,

Maddix²,

Aubet³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Quantile regression is an effective technique to quantify uncertainty, fit challenging underlying distributions, and often provide full probabilistic predictions through joint learnings over multiple quantile levels. A common drawback of these joint quantile regressions, however, is quantile crossing, which violates the desirable monotone property of the conditional quantile function. In this work, we propose the Incremental (Spline) Quantile Functions I(S)QF, a flexible and efficient distribution-free quantile estimation framework that resolves quantile crossing with a simple neural network layer. Moreover, I(S)QF inter/extrapolate to predict arbitrary quantile levels that differ from the underlying training ones. Equipped with the analytical evaluation of the continuous ranked probability score of I(S)QF representations, we apply our methods to NN-based times series forecasting cases, where the savings of the expensive re-training costs for non-trained quantile levels is particularly significant. We also provide a generalization error analysis of our proposed approaches under the sequence-tosequence setting. Lastly, extensive experiments demonstrate the improvement of consistency and accuracy errors over other baselines.

show abstract

“…We implement the model in the open-source GluonTS 4 [1] library. The experiments are done using AWS Sage-Maker [17].…”

Section: Methodsmentioning

confidence: 99%

Learning Quantile Functions without Quantile Crossing for Distribution-free Time Series Forecasting

Park¹,

Maddix²,

Aubet³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…We only use the root-level FOS for each paper and there are 19 such FOS. We use Amazon Sagemaker (Liberty et al, 2020) to run our experiments. For simplicity, we assume that the propensities π ij depend only on the FOS of papers p i and p j .…”

Section: Methodsmentioning

confidence: 99%

Correcting Exposure Bias for Link Recommendation

Gupta,

Wang,

Lipton

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Link prediction methods are frequently applied in recommender systems, e.g., to suggest citations for academic papers or friends in social networks. However, exposure bias can arise when users are systematically underexposed to certain relevant items. For example, in citation networks, authors might be more likely to encounter papers from their own field and thus cite them preferentially. This bias can propagate through naively trained link predictors, leading to both biased evaluation and high generalization error (as assessed by true relevance). Moreover, this bias can be exacerbated by feedback loops. We propose estimators that leverage known exposure probabilities to mitigate this bias and consequent feedback loops. Next, we provide a loss function for learning the exposure probabilities from data. Finally, experiments on semi-synthetic data based on real-world citation networks, show that our methods reliably identify (truly) relevant citations. Additionally, our methods lead to greater diversity in the recommended papers' fields of study. The code is available at github.com/shantanu95/ exposure-bias-link-rec.

show abstract

“…It follows POP's scheduling algorithm which combines probabilistic model-based classication with dynamic scheduling and early stop techniques. Amazon SageMaker [36] is a fully managed machine learning service. It supports automatic model tuning component that nds the best version of a model by running many training trials on the dataset using the algorithm and ranges of hyperparameters specied by the user.…”

Section: Related Workmentioning

confidence: 99%

“…We then observe that the cost of doing so grows exponentially with the number of parameters being tuned, becoming impractical. Commercial platforms (i.e., Google Vizier [19], Amazon SageMaker [36]), as well as on-premises solutions (i.e., Auto-Keras [25]) help deployers by oering tuning services to mitigate (possibly avoid) misconguration.…”

Section: Introductionmentioning

confidence: 99%