Papers › SurgeryV2: Bridging the Gap Between Model Merging and Multi-Task Learning with Deep...

SurgeryV2: Bridging the Gap Between Model Merging and Multi-Task Learning with Deep Representation Surgery

18 Oct 2024arXiv:2410.14389archive 2025-07-28

Enneng Yang, Li Shen, Zhenyi Wang, Guibing Guo, Xingwei Wang, Xiaocun Cao, Jie Zhang, DaCheng Tao

Model merging-based multitask learning (MTL) offers a promising approach for performing MTL by merging multiple expert models without requiring access to raw training data. However, in this paper, we examine the merged model's representation distribution and uncover a critical issue of "representation bias". This bias arises from a significant distribution gap between the representations of the merged and expert models, leading to the suboptimal performance of the merged MTL model. To address this challenge, we first propose a representation surgery solution called Surgery. Surgery is a lightweight, task-specific module that aligns the final layer representations of the merged model with those of the expert models, effectively alleviating bias and improving the merged model's performance. Despite these improvements, a performance gap remains compared to the traditional MTL method. Further analysis reveals that representation bias phenomena exist at each layer of the merged model, and aligning representations only in the last layer is insufficient for fully reducing systemic bias because biases introduced at each layer can accumulate and interact in complex ways. To tackle this, we then propose a more comprehensive solution, deep representation surgery (also called SurgeryV2), which mitigates representation bias across all layers, and thus bridges the performance gap between model merging-based MTL and traditional MTL. Finally, we design an unsupervised optimization objective to optimize both the Surgery and SurgeryV2 modules. Our experimental results show that incorporating these modules into state-of-the-art (SOTA) model merging schemes leads to significant performance gains. Notably, our SurgeryV2 scheme reaches almost the same level as individual expert models or the traditional MTL model. The code is available at https://github.com/EnnengYang/SurgeryV2.

PaperPDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 12 code samples harvested from 1 repository linked to this paper: 9 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 0 violated, 3 with no contract checked; 5 where Syntology's instrument failed) · 3 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 12 samples from 1 repository, 9 ran (4 with no instrument failure; 5 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 9 samples that ran have no recorded date.

ennengyang/surgeryv2 officialmentioned in papermentioned on GitHubpytorchSyntology: 12 samples, 9 ran (5 instrument-failed) · 3 unverified 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

12 samples harvested: 9 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 1 honoured, 0 violated, 3 with no contract checked; 5 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
5ran · our draft was wrong
3ran
3unverified

Licence: 12 of the 12 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 ennengyang/surgeryv2. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

ennengyang/surgeryv2 official repository 12 samples, 9 ran (5 instrument-failed) · 3 unverified
cosine_lr in 31 papers src/utils.py ran no licence file found · pointer only · 2d5d7a12d8de02b7 · report fixture: signature optimizer:SGD/0/?, base_lrs:list/1/float, warmup_length:int/0/int, steps:int/0/int · returned function · level recorded 2026-09-24
get_merging_cofficients src/merging_cofficient.py ran no licence file found · pointer only · f52df35accf87444 · report fixture: signature method:str/0/str, model:str/0/str, task_num:int/0/int · returned list · level recorded 2026-09-24
l1losses src/merging_model.py ran · our draft was wrong no licence file found · pointer only · 2ca594138c45f6d6 · report fixture: signature predicts:list/1/Tensor, labels:list/1/Tensor · returned Tensor · level recorded 2026-09-24
make_functional in 10 papers src/merging_model.py ran · our draft was wrong no licence file found · pointer only · 8adbe30621fbeb18 · report fixture: signature mod:callable/0/? · returned tuple · level recorded, date not recorded
softmax_entropy in 54 papers
census: 18 implementations from 29 papers, 4 distinct outputs across 2 buckets, one shared input per bucket
src/merging_model.py ran · our draft was wrong fingerprinted no licence file found · pointer only · 0c501eba6aaad5fa · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-24
split_train_into_train_val src/datasets/registry.py ran · our draft was wrong no licence file found · pointer only · 1e9d453dc1605e8c · report fixture: signature dataset:GenericDataset/0/?, new_dataset_class_name:str/0/str, batch_size:int/0/int, num_workers:int/0/int, val_fraction:float/0/float, max_val_samples:NoneType/0/?, seed:int/0/int · returned TestDataset · level recorded, date not recorded
state_dict_to_vector in 6 papers src/ties_merging_utils.py ran · honoured contract no licence file found · pointer only · 72391ca34e9a8cfe · report fixture: signature state_dict:dict/0/?, remove_keys:list/1/str · returned Tensor · level recorded 2026-09-22
vector_to_state_dict in 5 papers src/ties_merging_utils.py ran · our draft was wrong no licence file found · pointer only · bf6efec91354af72 · report fixture: signature vector:Tensor/1/float, state_dict:dict/0/?, remove_keys:list/1/str · returned OrderedDict · level recorded 2026-09-22
vector_to_state_dict_unsort src/ties_merging_utils.py ran no licence file found · pointer only · 2b9cb736cd2df186 · report fixture: signature vector:Tensor/1/float, state_dict:dict/0/?, remove_keys:list/1/str · returned dict · level recorded 2026-09-24
ModelWrapper_Surgery_V2 src/merging_model.py unverified no licence file found · pointer only · 539d129135c1c4c8 · report
accuracy in 872 papers
census: 8 implementations from 7 papers, 7 distinct outputs across 6 buckets, one shared input per bucket
src/utils.py unverified no licence file found · pointer only · aafcff3ad1ccef6e · report
torch_load_old in 8 papers src/utils.py unverified no licence file found · pointer only · 7d051a7adefedb6c · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly shared with arXiv:2402.02705: its method name (Surgery) and claim were written for that paper, not this one; Syntology's graph stores one assembly node for identical code attached to both. 1 brick; individually on a fixture: 0 ran · 1 unverified. The assembly has not been run end-to-end.

  1. ModelWrapper_Surgery_V2 entry · level 0 · unverified · 539d129135c1c4c8

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2410.14389") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Multi-Task Learning

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

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