Paper Citation Record · LEDGER

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining

As of 14 August 2026, this Paper Citation Record lists 35 of 35 outbound references and 0 inbound Pith citation observations for arXiv:2605.07756.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2605.07756 v1

Coverage vector

measured 35 of 35 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-05-11T03:11:37.361024Z

measured 35 of 35 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-14T06:32:32.682623+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: cited_works

Reference resolution

35 of 35 outbound references displayed

verified exact6
verified fuzzy28
unresolved1
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation d552e6eb-9a94-42b1-b84b-9fb6655d1583 · outbound

This paper cites Bert: Pre-training of deep bidirectional transformers for language understanding.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Bert: Pre-training of deep bidirectional transformers for language understanding

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.607229Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation cf5a6a5c-7fdd-49bf-bf04-257108310200 · outbound

This paper cites Bootstrap your own latent-a new approach to self-supervised learning.Advances in neural information processing systems, 33:21271–21284.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Bootstrap your own latent-a new approach to self-supervised learning.Advances in neural information processing systems, 33:21271–21284

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.614672Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 56ccf934-d434-4b2b-88ee-c0c6126b5af4 · outbound

This paper cites Coles: Contrastive learning for event sequences with self-supervision.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Coles: Contrastive learning for event sequences with self-supervision

Reference 3

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.603500Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 4254882a-b524-4a39-a91c-b1e8a1173e36 · outbound

This paper cites Tabular transformers for modeling multivariate time series.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Tabular transformers for modeling multivariate time series

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.612967Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation c78efa33-0afd-46d8-aa2c-96287fbbcef3 · outbound

This paper cites All4one: Symbiotic neighbour contrastive learning via self-attention and redundancy reduction.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining All4one: Symbiotic neighbour contrastive learning via self-attention and redundancy reduction

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.609763Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation ddfc6634-046a-4e89-9be8-2b27a5f58115 · outbound

This paper cites HT-Transformer: Event Sequences Classification by Accumulating Prefix Information with History Tokens.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining HT-Transformer: Event Sequences Classification by Accumulating Prefix Information with History Tokens

Reference 6

Resolution: verified exact; arxiv_id, observed 2026-05-11T03:15:54.998290Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 467cc0b5-f444-4439-9f6b-f8bb1f054f87 · outbound

This paper cites Practical bayesian optimization of machine learning algorithms.Advances in neural information processing systems, 25.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Practical bayesian optimization of machine learning algorithms.Advances in neural information processing systems, 25

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.605170Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 30497b99-bf06-4e7a-a238-7eddab3750d4 · outbound

This paper cites Gradient surgery for multi-task learning.Advances in neural information processing systems, 33:5824–5836.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Gradient surgery for multi-task learning.Advances in neural information processing systems, 33:5824–5836

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.616419Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 207251bf-82e1-45e7-9586-dbb3a5bb343d · outbound

This paper cites Bilevel programming for hyperparameter optimization and meta-learning.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Bilevel programming for hyperparameter optimization and meta-learning

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.611401Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation a3643a71-072c-407a-a83d-4df08ecedf41 · outbound

This paper cites Multi-task learning as multi-objective optimization.Advances in neural information processing systems, 31.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Multi-task learning as multi-objective optimization.Advances in neural information processing systems, 31

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.566750Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 2fd20c1d-17dd-4109-8d68-93342163071b · outbound

This paper cites Truncated back- propagation for bilevel optimization.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Truncated back- propagation for bilevel optimization

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.568587Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 00e0887f-1d2d-42ad-a688-a9c44e1c2d50 · outbound

This paper cites Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Gradnorm: Gradient normalization for adaptive loss balancing in deep multitask networks

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.570401Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 63299872-af95-4716-b5af-521f26472cd4 · outbound

This paper cites Independent component alignment for multi-task learning.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Independent component alignment for multi-task learning

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.572112Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation cdadf241-a899-40d6-85e2-43efce0d9fbf · outbound

This paper cites End-to-end multi-task learning with attention.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining End-to-end multi-task learning with attention

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.573832Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation f04a78a6-887e-4b4b-9049-f71460c693d4 · outbound

This paper cites Multi-task learning using uncertainty to weigh losses for scene geometry and semantics.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Multi-task learning using uncertainty to weigh losses for scene geometry and semantics

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.591220Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation bf9d8ccf-064a-4c70-a701-9d266f4c8893 · outbound

This paper cites Pytorch-lifestream: Learning embeddings on discrete event sequences.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Pytorch-lifestream: Learning embeddings on discrete event sequences

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.592840Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 20f35488-f6f4-4a60-97f3-1eb849ff8ca1 · outbound

This paper cites Optuna: A next-generation hyperparameter optimization framework.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Optuna: A next-generation hyperparameter optimization framework

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.594670Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 9d3938a6-60df-4553-a6bc-71001d17775f · outbound

This paper cites solo- learn: A library of self-supervised methods for visual representation learning.Journal of Machine Learning Research, 23(56):1–6.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining solo- learn: A library of self-supervised methods for visual representation learning.Journal of Machine Learning Research, 23(56):1–6

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.596380Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation f7f7cd3c-e0ea-4bfb-9317-58efbe75ca74 · outbound

This paper cites Learning multiple layers of features from tiny images.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Learning multiple layers of features from tiny images

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.601881Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation c07e6367-aca1-4246-ade4-6de145498799 · outbound

This paper cites Imagenet classification with deep convolutional neural networks.Advances in neural information processing systems, 25.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Imagenet classification with deep convolutional neural networks.Advances in neural information processing systems, 25

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.598173Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 07e9d31c-9f34-4855-a3db-b4b551199231 · outbound

This paper cites Barlow twins: Self- supervised learning via redundancy reduction.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Barlow twins: Self- supervised learning via redundancy reduction

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.575742Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 868ec518-b03a-4875-ad7b-d8637428dfba · outbound

This paper cites With a little help from my friends: Nearest-neighbor contrastive learning of visual represen- tations.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining With a little help from my friends: Nearest-neighbor contrastive learning of visual represen- tations

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.577408Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation cb27ac88-cc26-4963-b27c-645d9ddec4c3 · outbound

This paper cites Making a science of model search: Hyper- parameter optimization in hundreds of dimensions for vision architectures.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Making a science of model search: Hyper- parameter optimization in hundreds of dimensions for vision architectures

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.599951Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 02b22c3e-f36c-4f72-a4b2-b5f94d8b47af · outbound

This paper cites an unresolved cited work.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Unresolved cited work

Reference 24

Resolution: unresolved; raw_fallback, observed 2026-05-14T10:54:13.589540Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation ad269814-27ac-4ed8-bdfd-65802e223286 · outbound

This paper cites DrMAD: Distilling Reverse-Mode Automatic Differentiation for Optimizing Hyperparameters of Deep Neural Networks.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining DrMAD: Distilling Reverse-Mode Automatic Differentiation for Optimizing Hyperparameters of Deep Neural Networks

Reference 25

Resolution: verified exact; arxiv_id, observed 2026-07-04T20:56:49.839258Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation dbbff8e9-48df-47f8-97ee-d748351907ba · outbound

This paper cites Scalable gradient-based tuning of continuous regularization hyperparameters.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Scalable gradient-based tuning of continuous regularization hyperparameters

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.582617Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation b55e35eb-3421-4bac-ac50-901688a1d78f · outbound

This paper cites Optimizing millions of hyperparameters by implicit differentiation.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Optimizing millions of hyperparameters by implicit differentiation

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.584169Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 7b7d486c-8c12-468b-8535-fa52955b9979 · outbound

This paper cites Jacobian Descent for Multi-Objective Optimization.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Jacobian Descent for Multi-Objective Optimization

Reference 28

Resolution: verified exact; arxiv_id, observed 2026-05-11T03:15:55.022903Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 3ed88404-e0f5-4a8b-822e-f83ffb33990d · outbound

This paper cites Unpreju- diced training auxiliary tasks makes primary better: A multitask learning perspective.IEEE Transactions on Neural Networks and Learning Systems, 36(7):12091–12105.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Unpreju- diced training auxiliary tasks makes primary better: A multitask learning perspective.IEEE Transactions on Neural Networks and Learning Systems, 36(7):12091–12105

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.586063Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 0f74a29c-1303-41f3-be3d-89f3ca3b4493 · outbound

This paper cites Sample-level weighting for multi-task learning with auxiliary tasks: E.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Sample-level weighting for multi-task learning with auxiliary tasks: E

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.579138Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 0c60dbb2-a7cc-47f9-8246-07777fb4facb · outbound

This paper cites Ldc-mtl: Balancing multi-task learning through scalable loss discrepancy control.arXiv preprint arXiv:2502.08585.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Ldc-mtl: Balancing multi-task learning through scalable loss discrepancy control.arXiv preprint arXiv:2502.08585

Reference 31

Resolution: verified exact; arxiv_id, observed 2026-05-11T03:15:55.016588Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 70316813-0b1e-4947-ac4e-0fdb89d36d78 · outbound

This paper cites Adapting Auxiliary Losses Using Gradient Similarity.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Adapting Auxiliary Losses Using Gradient Similarity

Reference 32

Resolution: verified exact; arxiv_id, observed 2026-05-11T03:15:54.976625Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 7515c056-cec7-4240-a474-e02ae21a7d3a · outbound

This paper cites Adaptive auxiliary task weighting for reinforcement learning.Advances in neural information processing systems, 32.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Adaptive auxiliary task weighting for reinforcement learning.Advances in neural information processing systems, 32

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.581058Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 5d287b56-00e7-41e7-b5ae-33d9f74398b2 · outbound

This paper cites Adam: A method for stochastic optimization.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Adam: A method for stochastic optimization

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-05-14T10:54:13.587909Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Observation 05852816-1d02-49e7-9687-aca73fbcff49 · outbound

This paper cites Large Batch Training of Convolutional Networks.

When Losses Align: Gradient-Based Composite Loss Weighting for Efficient Pretraining Large Batch Training of Convolutional Networks

Reference 35

Resolution: verified exact; arxiv_id, observed 2026-05-11T03:15:54.991024Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-14T06:32:32.682623+00:00.

Pith citing papers

No inbound Pith citation observations are available.