# isl-org/MultiObjectiveOptimization

Source code for Neural Information Processing Systems (NeurIPS) 2018 paper "Multi-Task Learning as Multi-Objective Optimization"

Repository: https://github.com/isl-org/MultiObjectiveOptimization
Canonical: https://ross.abutalabs.com/products/multiobjectiveoptimization
Language: Python
License: MIT
License Family: permissive
Archived: true
Last push: 2024-09-02T19:26:42+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2778, "days_push": 730, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1070, forks 174 (observed 2026-08-28T04:03:28.010468+00:00)

## What it is
Source code for the NeurIPS 2018 paper 'Multi-Task Learning as Multi-Objective Optimization', implementing the MGDA_UB algorithm for multi-task learning. It includes PyTorch and pure NumPy implementations of Frank-Wolfe and projected gradient descent solvers, plus the MultiMNIST dataset.

## Use cases
- implement multi-task learning as multi-objective optimization
- balance gradients across multiple loss functions in deep learning
- reproduce the MGDA_UB algorithm from the NeurIPS 2018 paper
- train models on the MultiMNIST dataset
- use Frank-Wolfe or projected gradient descent for Pareto optimization
- port gradient surgery techniques to other deep learning frameworks

## When to choose
- you need the reference implementation of MGDA_UB for research or citation
- you want a framework-agnostic NumPy multi-objective gradient solver
- you are studying multi-task learning optimization methods

## When to avoid
- you need maintained software with bug fixes or updates - Intel has discontinued the project
- you need production-ready multi-task learning tooling with active community support
- you require support for recent PyTorch versions without your own patches

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, deep-learning
- domain: machine-learning, deep-learning
- platform: python, cross-platform
- tags: multi-task-learning, multi-objective-optimization, mgda, gradient-descent, pytorch, neurips-2018, research-code, numpy, optimization, research

## Member repositories
- isl-org/MultiObjectiveOptimization (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:28.010468+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:54:02.376122+00:00, confidence not recorded.
  - readme: https://github.com/isl-org/MultiObjectiveOptimization (fetched 2026-08-28T04:03:28.010468+00:00, sha dc518480eebe)
- Data as of 2026-08-30T08:39:29.467469+00:00.