Papers › DSelect-k: Differentiable Selection in the Mixture of Experts with Applications to...

DSelect-k: Differentiable Selection in the Mixture of Experts with Applications to Multi-Task Learning

7 Jun 2021NeurIPS 2021 12arXiv:2106.03760archive 2025-07-28

Hussein Hazimeh, Zhe Zhao, Aakanksha Chowdhery, Maheswaran Sathiamoorthy, Yihua Chen, Rahul Mazumder, Lichan Hong, Ed H. Chi

The Mixture-of-Experts (MoE) architecture is showing promising results in improving parameter sharing in multi-task learning (MTL) and in scaling high-capacity neural networks. State-of-the-art MoE models use a trainable sparse gate to select a subset of the experts for each input example. While conceptually appealing, existing sparse gates, such as Top-k, are not smooth. The lack of smoothness can lead to convergence and statistical performance issues when training with gradient-based methods. In this paper, we develop DSelect-k: a continuously differentiable and sparse gate for MoE, based on a novel binary encoding formulation. The gate can be trained using first-order methods, such as stochastic gradient descent, and offers explicit control over the number of experts to select. We demonstrate the effectiveness of DSelect-k on both synthetic and real MTL datasets with up to $128$ tasks. Our experiments indicate that DSelect-k can achieve statistically significant improvements in prediction and expert selection over popular MoE gates. Notably, on a real-world, large-scale recommender system, DSelect-k achieves over 22% improvement in predictive performance compared to Top-k. We provide an open-source implementation of DSelect-k.

PaperPDFConference PDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2106.03760")

Code

Syntology Ran 1 of 9 code samples harvested from 2 repositories linked to this paper; 8 have no recorded run. Of those that ran: 1 ran with no contract checked.

By repository: official repository: 2 samples from 1 repository, 1 ran; community (archive-listed): 7 samples from 1 repository, 0 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Andy1314Chen/PaddleRec paddleApache-2.0 report
PaddlePaddle/PaddleRec paddleApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

9 samples harvested; 1 ran; 0 honoured the contract we drafted; 8 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran
8unverified

Licence: 0 of the 9 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; each sample names its own and says which. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

SmoothStep google-research/google-research/dselect_k_moe/dselect_k_moe.py official repository ran fingerprinted Apache-2.0 (permissive) · c7f7716bc175db79 · report
DSelectKGate google-research/google-research/dselect_k_moe/dselect_k_moe.py official repository unverified Apache-2.0 (permissive) · 7f0363937bd725ba · report
fun1 Andy1314Chen/PaddleRec/datasets/census/data_preparation.py community (archive-listed) unverified Apache-2.0 (permissive) · b0a4669173a5fc3d · report
fun2 Andy1314Chen/PaddleRec/datasets/census/data_preparation.py community (archive-listed) unverified Apache-2.0 (permissive) · 4b20e0a2441ecb28 · report
load_graph Andy1314Chen/PaddleRec/datasets/AmazonBook/preprocess.py community (archive-listed) unverified Apache-2.0 (permissive) · e192b64a7c6ac2de · report
native_to_unicode Andy1314Chen/PaddleRec/datasets/one_billion/preprocess.py community (archive-listed) unverified Apache-2.0 (permissive) · 30082085af20d54d · report
read_embedding Andy1314Chen/PaddleRec/datasets/letor07/process.py community (archive-listed) unverified Apache-2.0 (permissive) · d1c5bda937752b7c · report
text_strip Andy1314Chen/PaddleRec/datasets/one_billion/preprocess.py community (archive-listed) unverified Apache-2.0 (permissive) · fc15fd4903900e23 · report
word_count Andy1314Chen/PaddleRec/datasets/ag_news/text2paddle.py community (archive-listed) unverified Apache-2.0 (permissive) · 5d079a1802fd6cb9 · report

Tasks

Mixture-of-ExpertsMulti-Task LearningRecommendation Systems

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Introduced by this paper: DSelect-k

DSelect-k

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections