# yields

Published articles for yields.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Guided Star-Shaped Masked Diffusion

DevFeed: [Guided Star-Shaped Masked Diffusion](<https://devfeed.tech/articles/guided-star-shaped-masked-diffusion-123072.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2510.08369>)

Author: Viacheslav Meshchaninov, Egor Shibaev, Artem Makoian, Ivan Klimov, Nikita Balagansky, Daniil Gavrilov, Aibek Alanov, Dmitry Vetrov

Published: 2026-11-09T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [Enhancing AI model performance](<https://devfeed.tech/topics/enhancing-ai-model-performance.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>)

Tags: [algorithm](<https://devfeed.tech/tags/algorithm.md>), [code-generation](<https://devfeed.tech/tags/code-generation.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [diffusion](<https://devfeed.tech/tags/diffusion.md>), [diffusion-models](<https://devfeed.tech/tags/diffusion-models.md>), [efficiency](<https://devfeed.tech/tags/efficiency.md>), [error](<https://devfeed.tech/tags/error.md>), [error-correction](<https://devfeed.tech/tags/error-correction.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [generation](<https://devfeed.tech/tags/generation.md>), [layer](<https://devfeed.tech/tags/layer.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [masking](<https://devfeed.tech/tags/masking.md>), [novel](<https://devfeed.tech/tags/novel.md>), [quality](<https://devfeed.tech/tags/quality.md>), [sampling](<https://devfeed.tech/tags/sampling.md>), [scheduler](<https://devfeed.tech/tags/scheduler.md>), [text](<https://devfeed.tech/tags/text.md>), [usability](<https://devfeed.tech/tags/usability.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

The proposed sampling algorithm aims to address irreversible decisions and weak low-step performance in pre-trained masked diffusion models. It reorganizes generation into a star-shaped process that permits error correction and uses a learnable re-masking scheduler to identify and revise likely errors. The method requires lightweight fine-tuning of a single layer. The abstract reports improved sample quality and efficiency, especially with few sampling steps, and says experiments on text and code generation outperform or match existing methods; it provides no specific metrics in the supplied excerpt.

### Source excerpt

arXiv:2510.08369v3 Announce Type: replace Abstract: The performance of pre-trained masked diffusion models is often constrained by their sampling procedure, which makes decisions irreversible and struggles in low-step generation regimes. We introduce a novel sampling algorithm that works with pre-trained models and, after a lightweight fine-tuning of a single layer, significantly improves sample quality and efficiency. Our method reformulates the generation process using a star-shaped paradigm, which inherently allows for error correction. To make this process effective, we augment it with a learnable re-masking scheduler that intelligently identifies and revises likely errors. This approach yields a substantial quality boost, particularly when using a small number of sampling steps. We extensively ablate key components of our approach and show its usability in different scenarios. In comprehensive experiments on text, and code generation, our sampling algorithm outperforms or matches existing methods.

## Benchmarking Unsupervised Jamming Detection in 5G Open RAN

DevFeed: [Benchmarking Unsupervised Jamming Detection in 5G Open RAN](<https://devfeed.tech/articles/benchmarking-unsupervised-jamming-detection-in-5g-open-ran-116934.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.04369>)

Author: Sneihil Gopal, Fernando J. Cintron, Richard A. Rouil

Published: 2026-11-04T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [5G](<https://devfeed.tech/topics/5g.md>), [threat detection](<https://devfeed.tech/topics/threat-detection.md>), [Ground truth / benchmark quality](<https://devfeed.tech/topics/ground-truth-benchmark-quality.md>)

Tags: [5g](<https://devfeed.tech/tags/5g.md>), [attack](<https://devfeed.tech/tags/attack.md>), [autoencoder](<https://devfeed.tech/tags/autoencoder.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [coherence](<https://devfeed.tech/tags/coherence.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [detection](<https://devfeed.tech/tags/detection.md>), [environments](<https://devfeed.tech/tags/environments.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [long-short-term-memory](<https://devfeed.tech/tags/long-short-term-memory.md>), [lstm](<https://devfeed.tech/tags/lstm.md>), [open](<https://devfeed.tech/tags/open.md>), [telemetry](<https://devfeed.tech/tags/telemetry.md>), [vae](<https://devfeed.tech/tags/vae.md>), [validation](<https://devfeed.tech/tags/validation.md>), [variable](<https://devfeed.tech/tags/variable.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

The benchmark finds that a Statistical Ensemble Method (SEM) detector outperformed Vanilla, LSTM, and VAE reconstruction-based baselines on KPI telemetry from ten heterogeneous 5G environments. In leave-one-out evaluation, SEM scored 0.906 on a composite measure combining detection confidence and temporal interval coherence, versus 0.688 for the best autoencoder per fold; ground-truth validation gave SEM a mean F1 of 0.918 versus 0.521 for the best AE variant. Composite rankings matched F1 rankings across environments, but the measure failed in within-trace and cross-trace evaluations when detectors flagged far fewer samples than the jamming prevalence. The authors therefore qualify its ranking reliability to cases where detection rate matches expected prevalence.

### Source excerpt

arXiv:2611.04369v1 Announce Type: cross Abstract: Radio Frequency (RF) jamming threatens fifth-generation (5G) networks. Open Radio Access Network (Open RAN) telemetry enables jamming detection without dedicated spectrum monitoring equipment, but labeled attack data are rarely available across deployments, motivating unsupervised detectors. Comparing such detectors across algorithmic families, however, remains difficult without ground-truth labels. This paper benchmarks statistical and reconstruction-based unsupervised jamming detectors using a symmetric evaluation interface that combines detection confidence and temporal interval coherence into a detector-independent composite score. We compare a Statistical Ensemble Method (SEM) detector against Vanilla, Long Short-Term Memory (LSTM), and Variational Autoencoder (VAE) baselines on Key Performance Indicator (KPI) telemetry from ten heterogeneous 5G environments with fixed and variable jamming bursts and dynamic jammer gains. Under leave-one-out (LOO) evaluation, SEM achieves a mean composite score of 0.906, compared with 0.688 for the best autoencoder (AE) per fold. Ground-truth validation yields a mean F1-score of 0.918 for SEM, compared with 0.521 for the best AE variant. The composite ranks SEM first in all environments, consistent with F1 rankings. However, in both within-trace and cross-trace evaluation, it fails for detectors that flag far fewer samples than the jamming prevalence, indicating that rankings are reliable only when detection rate matches the expected prevalence.

## Cheap Retrieval and Observation Credit Assignment from One Quantity: The Value of Information

DevFeed: [Cheap Retrieval and Observation Credit Assignment from One Quantity: The Value of Information](<https://devfeed.tech/articles/cheap-retrieval-and-observation-credit-assignment-from-one-quantity-the-value-of-information-108256.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.40030>)

Author: Daniel Strauss, Nikolaos Sotirakis, Zhenshan Bing, Alois Knoll

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [World models](<https://devfeed.tech/topics/world-models.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [information retrieval](<https://devfeed.tech/topics/information-retrieval.md>)

Tags: [attention](<https://devfeed.tech/tags/attention.md>), [context](<https://devfeed.tech/tags/context.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [energy](<https://devfeed.tech/tags/energy.md>), [latency](<https://devfeed.tech/tags/latency.md>), [lower-bound](<https://devfeed.tech/tags/lower-bound.md>), [memory](<https://devfeed.tech/tags/memory.md>), [model](<https://devfeed.tech/tags/model.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [robots](<https://devfeed.tech/tags/robots.md>), [sequence](<https://devfeed.tech/tags/sequence.md>), [set](<https://devfeed.tech/tags/set.md>), [signal](<https://devfeed.tech/tags/signal.md>), [store](<https://devfeed.tech/tags/store.md>), [system](<https://devfeed.tech/tags/system.md>), [target](<https://devfeed.tech/tags/target.md>), [with](<https://devfeed.tech/tags/with.md>), [world](<https://devfeed.tech/tags/world.md>), [world-model](<https://devfeed.tech/tags/world-model.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

The paper proposes using value of information, estimated from archived rollouts with a learned world model, to address two reinforcement-learning memory problems. At an observation, this estimate gives a provable lower bound on the optimal value of the action that acquired the information, floors its TD target, and assigns credit to that observation. At a decision, the same quantity defines the optimal retrieval set. A pilot retrieval system using this signal solves a memory task with reduced context; the abstract does not report further evaluation details.

### Source excerpt

arXiv:2610.40030v1 Announce Type: new Abstract: Memory in reinforcement learning raises two problems. Attention finds the relevant item by scoring the whole store, which is costly for robots under latency and energy limits. When a retrieved memory pays off, credit for acquiring it must cross the whole sequence. We argue that both are governed by one quantity, the value of information, and propose to estimate it from archived rollouts with a learned world model. We obtain that, when read at the observation, it yields a provable lower bound on the optimal value of the acquiring action, which floors its TD target and attributes this floor to the observation that earned it. Read at the decision, it defines the optimal retrieval set. A pilot retrieval system, to be trained and improved with this signal, solves a memory task with reduced context.

## Latent Prediction Yields Cross-Subject Alignable Representations in Sensory-Evoked fMRI

DevFeed: [Latent Prediction Yields Cross-Subject Alignable Representations in Sensory-Evoked fMRI](<https://devfeed.tech/articles/latent-prediction-yields-cross-subject-alignable-representations-in-sensory-evoked-fmri-108283.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.40280>)

Author: Serwan Jassim, Victoria Bosch, Daniel Anthes, Elia Bruni, Tim Kietzmann

Published: 2026-10-30T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Representation learning](<https://devfeed.tech/topics/representation-learning.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Decoding](<https://devfeed.tech/topics/decoding.md>), [Medical imaging](<https://devfeed.tech/topics/medical-imaging.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [alignment](<https://devfeed.tech/tags/alignment.md>), [approach](<https://devfeed.tech/tags/approach.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [brain](<https://devfeed.tech/tags/brain.md>), [can](<https://devfeed.tech/tags/can.md>), [central](<https://devfeed.tech/tags/central.md>), [cross](<https://devfeed.tech/tags/cross.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [data](<https://devfeed.tech/tags/data.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [datasets](<https://devfeed.tech/tags/datasets.md>), [decoding](<https://devfeed.tech/tags/decoding.md>), [domain](<https://devfeed.tech/tags/domain.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [fine](<https://devfeed.tech/tags/fine.md>), [fine-tuning](<https://devfeed.tech/tags/fine-tuning.md>), [prediction](<https://devfeed.tech/tags/prediction.md>), [q-bio-nc](<https://devfeed.tech/tags/q-bio-nc.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

The study adapts a joint embedding predictive architecture (JEPA) to learn participant-specific fMRI representations from brain data alone, without stimulus images or derived representations. On the Natural Scenes Dataset, the embeddings support brain captioning and image reconstruction at competitive performance with parameter-rich supervised pipelines that use image and language data. Independently trained participant encoders can also be aligned without images or shared stimuli; after alignment, a decoding head trained on one participant transfers to another without fine-tuning and retains most within-subject performance on both tasks. The authors present this as a step toward pooling fMRI data across participants and datasets.

### Source excerpt

arXiv:2610.40280v1 Announce Type: cross Abstract: Decoding sensory-evoked functional magnetic resonance imaging (fMRI) data is a central analytical tool in cognitive neuroscience. Here we ask whether self-supervised learning (SSL), which has been scarcely explored in this domain, can yield representations that support strong downstream performance. To do so, we adapt a joint embedding predictive architecture (JEPA) to train participant-specific encoders on brain data alone. The encoders never see the stimulus images or any representation derived from them. Applied to the Natural Scenes Dataset (NSD), we evaluate our learned brain embeddings on the downstream tasks of brain captioning and image reconstruction, as well as cross-participant generalization. Our self-supervised approach achieves competitive performance with parameter-rich supervised pipelines that are trained end-to-end with access to image and language data. We further find that encoders trained independently for different participants can be aligned fully unsupervised, using neither images nor stimuli shared between them. After alignment, a neural decoding head trained on one participant can be applied to another without fine-tuning and retains most of its within-subject performance on image reconstruction and brain captioning. By removing the need for images and shared stimuli, our approach is a step towards pooling fMRI data across participants and, eventually, across datasets.

## DISC: Domain-Integrated Similarity Cost for Time-Series Dataset

DevFeed: [DISC: Domain-Integrated Similarity Cost for Time-Series Dataset](<https://devfeed.tech/articles/disc-domain-integrated-similarity-cost-for-time-series-dataset-97044.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.31288>)

Author: Yen-Ku Liu, Hongjie Chen, Franck Dernoncourt, Ryan Rossi

Published: 2026-10-23T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Time Series](<https://devfeed.tech/topics/time-series.md>), [AI Foundation Models](<https://devfeed.tech/topics/ai-foundation-models.md>)

Tags: [aggregation](<https://devfeed.tech/tags/aggregation.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [forecasting](<https://devfeed.tech/tags/forecasting.md>), [learning](<https://devfeed.tech/tags/learning.md>), [metric](<https://devfeed.tech/tags/metric.md>), [proxy](<https://devfeed.tech/tags/proxy.md>), [select](<https://devfeed.tech/tags/select.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [time-series](<https://devfeed.tech/tags/time-series.md>), [transfer-learning](<https://devfeed.tech/tags/transfer-learning.md>), [weight](<https://devfeed.tech/tags/weight.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

DISC combines time-series dataset similarity metrics through a learnable framework calibrated on transfer observations. It selects, weights, and integrates metrics for downstream tasks, then scores source-target pairs using their metric values without evaluating every pair through model transfer. In the reported out-of-distribution and transfer-learning split, its scores align better with downstream forecasting errors than single-metric references. The abstract presents similarity as task-dependent; it does not provide detailed evaluation results beyond that reported comparison.

### Source excerpt

arXiv:2610.31288v1 Announce Type: new Abstract: The growth of time-series datasets and foundation models has made dataset comparison a central part of time-series workflows. Assessing source-target transfer before model adaptation requires similarity estimates that can summarize transfer-related dataset relationships without exhaustive source-target evaluation over every candidate pair. Many metrics have been proposed for measuring similarity between time-series datasets, but they capture different dataset relationships and vary in how well they align with downstream utility across tasks. We evaluate this gap by measuring the correlation between source-target similarity scores and downstream forecasting errors in out-of-distribution and transfer-learning settings. We propose Domain-Integrated Similarity Cost (DISC), a learnable aggregation framework that uses calibration transfer observations to select, weight, and combine similarity metrics for different downstream tasks. By integrating complementary metric values, DISC better distinguishes source-target dataset differences that matter for downstream behavior. After calibration, DISC scores source-target pairs using only their similarity-metric values and yields higher alignment than single-metric references on the reported split. These results show that time-series dataset similarity is task-dependent, and that learned metric integration provides a stronger proxy for downstream transfer behavior than manually selecting one similarity metric.

## CoarseSoundNet: Building a reliable model for ecological soundscape analysis

DevFeed: [CoarseSoundNet: Building a reliable model for ecological soundscape analysis](<https://devfeed.tech/articles/coarsesoundnet-building-a-reliable-model-for-ecological-soundscape-analysis-96839.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2605.21143>)

Author: Alexander Gebhard, Andreas Triantafyllopoulos, Dominik Arend, Sandra M\\"uller, Svenja Schmidt, Michael Scherer-Lorenzen, Bj\\"orn W. Schuller

Published: 2026-10-23T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Model Development](<https://devfeed.tech/topics/model-development.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Data analysis](<https://devfeed.tech/topics/data-analysis.md>)

Tags: [case-study](<https://devfeed.tech/tags/case-study.md>), [challenges](<https://devfeed.tech/tags/challenges.md>), [classification](<https://devfeed.tech/tags/classification.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-sd](<https://devfeed.tech/tags/cs-sd.md>), [deep-learning](<https://devfeed.tech/tags/deep-learning.md>), [diversity](<https://devfeed.tech/tags/diversity.md>), [domain](<https://devfeed.tech/tags/domain.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [framework](<https://devfeed.tech/tags/framework.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [masking](<https://devfeed.tech/tags/masking.md>), [model](<https://devfeed.tech/tags/model.md>), [pam](<https://devfeed.tech/tags/pam.md>), [recordings](<https://devfeed.tech/tags/recordings.md>), [training](<https://devfeed.tech/tags/training.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

CoarseSoundNet classifies terrestrial passive acoustic monitoring recordings as biophony, geophony, or anthropophony. The study reports better performance with more PAM training data--especially data resembling the target domain--and with an explicit silence class. Post-hoc calibration using class-specific thresholds further improves performance; minimum annotation durations raise scores for anthropophony and geophony. Errors remain: masking challenges anthropophony classification, while silence and insect sounds are confused with geophony and biophony. In an ecological case study, filtering recordings with the model produced acoustic-index trends broadly similar to ground-truth filtering, but only limited improvement in relationships with avian alpha-diversity.

### Source excerpt

arXiv:2605.21143v3 Announce Type: replace-cross Abstract: A soundscape is composed of three types of sound: biophony (sounds made by animals), geophony (natural abiotic sounds) and anthropophony (sounds made by humans). A key research question in the field of soundscape ecology is how these components interact with each other, specifically how biophony responds to geophony and anthropophony. Nevertheless, as of today, there are not many analytical instruments that enable the distinct quantification of these elements. Recent machine learning (ML) approaches aim to support automated analysis but often rely on task-specific or clean data, limiting generalisation to noisy passive acoustic monitoring (PAM) recordings. This study presents a reproducible framework to build ML models for coarse soundscape classification and introduces CoarseSoundNet, a deep learning model trained to distinguish biophony, geophony, and anthropophony under realistic terrestrial PAM conditions. We systematically investigate model architectures, the influence of an additional training class, data composition, and evaluation strategies. Our findings suggest that model performance improves with additional PAM data, especially when similar to the target domain, and by introducing an explicit silence class during training. Post-hoc target-domain calibration employing class-specific decision thresholds further enhance performance, while incorporating minimum annotation durations yields higher scores for anthropophony and geophony. Error analyses exhibit challenges for anthropophony due to masking effects and confusions for silence and insect sounds for geophony and biophony. Finally, we conduct an ecological case study showing that pre-filtering recordings with CoarseSoundNet yields acoustic index trends broadly similar to ground-truth filtering, although the resulting improvements in their relationships with avian alpha-diversity were limited.

## Yields are Up, Latencies are Down: Goroutine Scheduling in CockroachDB

DevFeed: [Yields are Up, Latencies are Down: Goroutine Scheduling in CockroachDB](<https://devfeed.tech/articles/yields-are-up-latencies-are-down-goroutine-scheduling-in-cockroachdb-23789.md>)

Original publisher: [Read original article](<https://cockroachlabs.com/blog/goroutine-scheduling-elastic-admission-control-cockroachdb>)

Author: David Taylor

Published: 2026-04-14T00:00:00Z

Content type: article

Language: en

Sources: [Cockroach Labs](<https://devfeed.tech/sources/cockroach-labs.md>)

Topics: [scheduling](<https://devfeed.tech/topics/scheduling.md>), [CockroachDB](<https://devfeed.tech/topics/cockroachdb.md>), [Concurrency](<https://devfeed.tech/topics/concurrency.md>)

Tags: [background](<https://devfeed.tech/tags/background.md>), [background-work](<https://devfeed.tech/tags/background-work.md>), [backups](<https://devfeed.tech/tags/backups.md>), [cockroachdb](<https://devfeed.tech/tags/cockroachdb.md>), [goroutine](<https://devfeed.tech/tags/goroutine.md>), [scheduling](<https://devfeed.tech/tags/scheduling.md>), [yields](<https://devfeed.tech/tags/yields.md>)

### AI overview

CockroachDB v26.1 pairs elastic admission control with cooperative yielding in a patched Go runtime so background jobs can use CPU capacity while stepping aside when foreground work is ready. The controller adjusts admission using p99 scheduler latency; Yield checks the runtime's runnable work and parks background goroutines under saturation, using natural safe points such as loop iterations and batch boundaries. In reported tests, an IMPORT at 95%+ CPU utilization had scheduling delays below 200µs and concurrent YCSB latency peaked at 4ms, compared with prior delays of 30-50ms and latency spikes above 1000x. Yield overhead was under 1% during CPU-saturating bulk work. Yield complements rather than replaces admission control, which still throttles jobs if delays rise. Work with long critical sections or infrequent yield points can still cause delays, and adding yield points across the codebase remains ongoing.

### Source excerpt

CockroachDB runs background work -- backups, schema changes, statistics collection, changefeeds -- in the same processes that serve user queries. Traditionally, this creates a tension...