SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.09759 · ICML · 2023

Pareto Manifold Learning: Tackling multiple tasks via ensembles of single-task models

Pascal Frossard, Franc ¸ois Fleuret, Nikolaos Dimitriadis

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2210.09759.json · badge: /api/ran/2210.09759.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
nik-dim/pamal canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
FunctionStatusWhere it lives
PaMaL_GL Ran
ran on a synthesized input · constructed an object; no result was computed
nik-dim/pamal/src/methods/pamal_weight_methods.py
code served (permissive licence) · get_code("7f24b9b23b50ce5f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, PaMaL_GL: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Create a minimal shared model that has parameters
class SharedModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

shared_model = SharedModel()
shared_parameters = list(shared_model.parameters())

# Create losses tensor - 2 tasks, small values
losses = torch.tensor([0.5, 0.7], dtype=torch.float32, requires_grad=False)

# Create task weights tensor
task_weights = torch.tensor([1.0, 1.0], dtype=torch.float32)

# Number of tasks
num_tasks = 2

ARGS = {
    'num_tasks': num_tasks,
    'iteration_window': 25,
    'temp': 2.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type PaMaL_GL — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PaMaL_GL output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
PamalWeightMethod Ran
ran on a synthesized input · constructed an object; no result was computed
nik-dim/pamal/src/methods/pamal_weight_methods.py
code served (permissive licence) · get_code("87c368e49fbc3c8b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned an instance of its own class, PamalWeightMethod: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import logging

torch.manual_seed(0)

class PamalWeightMethod:
    def __init__(self, num_tasks: int):
        super().__init__()
        self.num_tasks = num_tasks
        logging.warn(f'Initializing {self.__repr__()}')

    def connect_device(self, device):
        self.device = device

    def get_weighted_loss(self, losses, shared_parameters, task_specific_parameters, last_shared_parameters, representation, **kwargs):
        weighted_loss = losses.mean()
        extra_outputs = {"info": "test"}
        return weighted_loss, extra_outputs

    def backward(self, losses, shared_parameters=None, task_specific_parameters=None, last_shared_parameters=None, representation=None, grad_scaler=None, **kwargs):
        loss, extra_outputs = self.get_weighted_loss(losses=losses, shared_parameters=shared_parameters, task_specific_parameters=task_specific_parameters, last_shared_parameters=last_shared_parameters, representation=representation, **kwargs)
        if grad_scaler:
            grad_scaler.scale(loss).backward()
        else:
            loss.backward()
        return (loss, extra_outputs)

    def __call__(self, losses, shared_parameters=None, task_specific_parameters=None, **kwargs):
        return self.backward(losses=losses, shared_parameters=shared_parameters, task_specific_parameters=task_specific_parameters, **kwargs)

    def parameters(self):
        return []


# Create a concrete subclass for testing
class ConcreteWeightMethod(PamalWeightMethod):
    def get_weighted_loss(self, losses, shared_parameters, task_specific_parameters, last_shared_parameters, representation, **kwargs):
        weighted_loss = losses.mean()
        extra_outputs = {"weighted": True}
        return weighted_loss, extra_outputs


num_tasks = 3
batch_size = 2
loss_dim = 4

# Create test data
losses = torch.randn(batch_size, loss_dim, requires_grad=True)
shared_params = [torch.randn(4, 3, requires_grad=True), torch.randn(3, 2, requires_grad=True)]
task_specific_params = [torch.randn(2, 4, requires_grad=True)]
last_shared_params = [torch.randn(5, 2, requires_grad=True)]
representation = torch.randn(batch_size, 4, requires_grad=True)

ARGS = {
    "num_tasks": num_tasks,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type PamalWeightMethod — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PamalWeightMethod output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

In Multi-Task Learning (MTL), tasks may compete and limit the performance achieved on each other, rather than guiding the optimization to a solution, superior to all its single-task trained counterparts. Since there is often not a unique solution optimal for all tasks, practitioners have to balance tradeoffs between tasks' performance, and resort to optimality in the Pareto sense. Most MTL methodologies either completely neglect this aspect, and instead of aiming at learning a Pareto Front, produce one solution predefined by their optimization schemes, or produce diverse but discrete solutions. Recent approaches parameterize the Pareto Front via neural networks, leading to complex mappings from tradeoff to objective space. In this paper, we conjecture that the Pareto Front admits a linear parameterization in parameter space, which leads us to propose Pareto Manifold Learning, an ensembling method in weight space. Our approach produces a continuous Pareto Front in a single training run, that allows to modulate the performance on each task during inference. Experiments on multitask learning benchmarks, ranging from image classification to tabular datasets and scene understanding, show that Pareto Manifold Learning outperforms state-of-the-art single-point algorithms, while learning a better Pareto parameterization than multi-point baselines.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2210.09759")
get_code_for_paper("2210.09759")
have("2210.09759")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2210.09759.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2210.09759.svg)](https://syntology.ai/paper/2210.09759)

Connect an agent — have() is free.