https://arxiv.org/abs/2409.12089

Skip to main content
Cornell University
We gratefully acknowledge support from the Simons Foundation, member
institutions, and all contributors. Donate
 
arxiv logo > cs > arXiv:2409.12089
[                    ]

Help | Advanced Search

[All fields        ]
Search
arXiv logo
Cornell University Logo
[                    ] GO
quick links

  * Login
  * Help Pages
  * About

Computer Science > Machine Learning

arXiv:2409.12089 (cs)
[Submitted on 18 Sep 2024 (v1), last revised 19 Sep 2024 (this
version, v2)]

Title:The Impact of Element Ordering on LM Agent Performance

Authors:Wayne Chi, Ameet Talwalkar, Chris Donahue
View a PDF of the paper titled The Impact of Element Ordering on LM
Agent Performance, by Wayne Chi and 2 other authors
View PDF HTML (experimental)

    Abstract:There has been a surge of interest in language model
    agents that can navigate virtual environments such as the web or
    desktop. To navigate such environments, agents benefit from
    information on the various elements (e.g., buttons, text, or
    images) present. It remains unclear which element attributes have
    the greatest impact on agent performance, especially in
    environments that only provide a graphical representation (i.e.,
    pixels). Here we find that the ordering in which elements are
    presented to the language model is surprisingly
    impactful--randomizing element ordering in a webpage degrades
    agent performance comparably to removing all visible text from an
    agent's state representation. While a webpage provides a
    hierarchical ordering of elements, there is no such ordering when
    parsing elements directly from pixels. Moreover, as tasks become
    more challenging and models more sophisticated, our experiments
    suggest that the impact of ordering increases. Finding an
    effective ordering is non-trivial. We investigate the impact of
    various element ordering methods in web and desktop environments.
    We find that dimensionality reduction provides a viable ordering
    for pixel-only environments. We train a UI element detection
    model to derive elements from pixels and apply our findings to an
    agent benchmark--OmniACT--where we only have access to pixels.
    Our method completes more than two times as many tasks on average
    relative to the previous state-of-the-art.

Subjects: Machine Learning (cs.LG)
Cite as:  arXiv:2409.12089 [cs.LG]
          (or arXiv:2409.12089v2 [cs.LG] for this version)
          https://doi.org/10.48550/arXiv.2409.12089
          Focus to learn more
          arXiv-issued DOI via DataCite

Submission history

From: Wayne Chi [view email]
[v1] Wed, 18 Sep 2024 16:04:10 UTC (3,177 KB)
[v2] Thu, 19 Sep 2024 05:44:21 UTC (3,177 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled The Impact of Element Ordering on
    LM Agent Performance, by Wayne Chi and 2 other authors
  * View PDF
  * HTML (experimental)
  * TeX Source
  * Other Formats

license icon view license
Current browse context:
cs.LG
< prev   |   next >
new | recent | 2024-09
Change to browse by:
cs

References & Citations

  * NASA ADS
  * Google Scholar
  * Semantic Scholar

a export BibTeX citation Loading...

BibTeX formatted citation

x
[loading...          ]
Data provided by:

Bookmark

BibSonomy logo Reddit logo
(*) Bibliographic Tools

Bibliographic and Citation Tools

[ ] Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
[ ] Litmaps Toggle
Litmaps (What is Litmaps?)
[ ] scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
( ) Code, Data, Media

Code, Data and Media Associated with this Article

[ ] Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
[ ] DagsHub Toggle
DagsHub (What is DagsHub?)
[ ] GotitPub Toggle
Gotit.pub (What is GotitPub?)
[ ] Links to Code Toggle
Papers with Code (What is Papers with Code?)
[ ] ScienceCast Toggle
ScienceCast (What is ScienceCast?)
( ) Demos

Demos

[ ] Replicate Toggle
Replicate (What is Replicate?)
[ ] Spaces Toggle
Hugging Face Spaces (What is Spaces?)
[ ] Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
( ) Related Papers

Recommenders and Search Tools

[ ] Link to Influence Flower
Influence Flower (What are Influence Flowers?)
[ ] Connected Papers Toggle
Connected Papers (What is Connected Papers?)
[ ] Core recommender toggle
CORE Recommender (What is CORE?)
[ ] IArxiv recommender toggle
IArxiv Recommender (What is IArxiv?)

  * Author
  * Venue
  * Institution
  * Topic

( ) About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and
share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have
embraced and accepted our values of openness, community, excellence,
and user data privacy. arXiv is committed to these values and only
works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community?
Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is
MathJax?)

  * About
  * Help

  * Click here to contact arXiv Contact
  * Click here to subscribe Subscribe

  * Copyright
  * Privacy Policy

  * Web Accessibility Assistance
  * arXiv Operational Status
    Get status notifications via email or slack