SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.13807 · 2024

AlanaVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 4 failures of our instrument; 0 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/2406.13807.json · badge: /api/ran/2406.13807.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
alanaai/evud canonical (classifier label)
classifier basis: census
4 ran · 0 no recorded run
FunctionStatusWhere it lives
generate_caption Ran
ran; our contract draft was wrong, not the code
alanaai/evud/hm3d/prepare_hm3d_dataset.py
code served (permissive licence) · get_code("d776259a0bbaf187")
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 str. The image it ran in was 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 random

def generate_caption(category_name):
    templates = ['The person is walking towards the {}.', 'The camera wearer is walking towards the {}.', "I'm walking over to the {} now.", "There's a person approaching the {}.", 'I see a person approaching the {}.', 'I see someone walking up to the {}.']
    return random.choice(templates).format(category_name)


ARGS = {
    'category_name': 'door'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
prepare_egoclip Ran
ran; our contract draft was wrong, not the code
alanaai/evud/egoclip/prepare_egoclip_dataset.py
code served (permissive licence) · get_code("182bf657bbecc36c")
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 DataFrame. The image it ran in was 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 random
import pandas as pd
import tempfile
import os

# Create a temporary CSV file with the expected structure
temp_dir = tempfile.mkdtemp()
csv_file = os.path.join(temp_dir, "egoclip_metadata.tsv")

# Create sample data matching the expected format
data = {
    'clip_start': [0, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55],
    'clip_end': [5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60],
    'clip_text': [
        '#C The person walks forward',
        '#O Another person sits down',
        'Summary Overall the video shows activity',
        '#C The camera wearer picks up an object',
        '#O Someone else opens a door',
        '#C Walking through the hallway',
        '#O Person waves hand',
        '#C Reaching for something on shelf',
        '#O #unsure Another person talking',
        '#C Looking at the camera',
        '#O Person running',
        '#C Turning around'
    ],
    'col1': ['a'] * 12,
    'col2': ['b'] * 12,
    'col3': ['c'] * 12,
    'col4': ['d'] * 12,
    'col5': ['e'] * 12,
    'col6': ['f'] * 12,
    'col7': ['g'] * 12,
    'col8': ['h'] * 12,
    'col9': ['i'] * 12,
    'Unnamed: 10': ['j'] * 12
}

df = pd.DataFrame(data)
df.to_csv(csv_file, sep='\t', index=False)

ARGS = {
    'egoclip_metadata': csv_file,
    'num_clips': 10,
    'min_duration': 2,
    'max_duration': 60
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
preprocess_data Ran
ran; our contract draft was wrong, not the code
alanaai/evud/gemini/prepare_ego4d_vqa_gemini_dataset.py
code served (permissive licence) · get_code("cd7e60a6f8674afa")
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 tuple. The image it ran in was 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 copy

def sort_dicts_by_category_list(dict_list, category_list):
    """Sorts a list of dictionaries based on the order of categories in a category list."""
    category_order = {category: i for i, category in enumerate(category_list)}
    sorted_dicts = sorted(dict_list, key=lambda item: category_order.get(item['category'], float('inf')))
    return sorted_dicts

def preprocess_data(gen_data):
    """Process the generated text - check how many responses fit the expected format and return the missing data too"""
    categories = ['object recognition', 'attribute recognition', 'object state recognition', 'object localisation', 'spatial reasoning', 'functional reasoning', 'world knowledge']
    processed_data = []
    missing = []
    for g in gen_data:
        try:
            if len(g['response']['candidates']) < 1:
                missing.append(g)
                continue
            if 'content' not in g['response']['candidates'][0]:
                missing.append(g)
                continue
            response = g['response']['candidates'][0]['content']['parts'][0]['text']
            category_data = response.split('\n\n')
            examples = []
            for c_data in category_data:
                c_lines = c_data.splitlines()
                c_lines = [c_line for c_line in c_lines if bool(c_line)]
                example = {}
                for c_line in c_lines:
                    key, value = c_line.split(': ', 1)
                    example[key.strip().lower()] = value.strip().lower().replace('localization', 'localisation')
                examples.append(example)
            sorted_examples = sort_dicts_by_category_list(examples, categories)
            for example, category in zip(sorted_examples, categories):
                assert category == example['category'].lower()
            pd = copy.deepcopy(g)
            pd['processed_examples'] = sorted_examples
            processed_data.append(pd)
        except Exception:
            print(g)
            raise
    if len(missing) > 0:
        print(f'There are {len(missing)} examples which do not have a Gemini response')
    return (processed_data, missing)


# Construct valid arguments for preprocess_data
gen_data = [
    {
        'id': 'test_1',
        'response': {
            'candidates': [
                {
                    'content': {
                        'parts': [
                            {
                                'text': 'Category: Object Recognition\nDescription: A cat\n\nCategory: Attribute Recognition\nDescription: Fluffy\n\nCategory: Object State Recognition\nDescription: Sitting\n\nCategory: Object Localisation\nDescription: Top left\n\nCategory: Spatial Reasoning\nDescription: Near window\n\nCategory: Functional Reasoning\nDescription: Playing\n\nCategory: World Knowledge\nDescription: Mammals'
                            }
                        ]
                    }
                }
            ]
        }
    }
]

ARGS = {
    'gen_data': gen_data
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is 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
sort_dicts_by_category_list Ran
ran; our contract draft was wrong, not the code
alanaai/evud/gemini/prepare_ego4d_vqa_gemini_dataset.py
code served (permissive licence) · get_code("6e9e9f43b2f4f37f")
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 list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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 numpy as np

# Construct arguments for sort_dicts_by_category_list
dict_list = [
    {'category': 'B', 'name': 'item1', 'value': 10},
    {'category': 'A', 'name': 'item2', 'value': 20},
    {'category': 'C', 'name': 'item3', 'value': 30},
    {'category': 'A', 'name': 'item4', 'value': 40},
]

category_list = ['A', 'B', 'C']

ARGS = {
    'dict_list': dict_list,
    'category_list': category_list,
}
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
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 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 4 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 0 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.

FunctionAlso present inEarliest-datedOthers named
generate_caption 1 other paper
1 same year, order not determinable
2024.findings-emnlp.649 (2024 (year only))
ALANAVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding
prepare_egoclip 1 other paper
1 same year, order not determinable
2024.findings-emnlp.649 (2024 (year only))
ALANAVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding
preprocess_data 1 other paper
1 same year, order not determinable
2024.findings-emnlp.649 (2024 (year only))
ALANAVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding
sort_dicts_by_category_list 1 other paper
1 same year, order not determinable
2024.findings-emnlp.649 (2024 (year only))
ALANAVLM: A Multimodal Embodied AI Foundation Model for Egocentric Video Understanding

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

AI personal assistants deployed via robots or wearables require embodied understanding to collaborate with humans effectively. However, current Vision-Language Models (VLMs) primarily focus on third-person view videos, neglecting the richness of egocentric perceptual experience. To address this gap, we propose three key contributions. First, we introduce the Egocentric Video Understanding Dataset (EVUD) for training VLMs on video captioning and question answering tasks specific to egocentric videos. Second, we present AlanaVLM, a 7B parameter VLM trained using parameter-efficient methods on EVUD. Finally, we evaluate AlanaVLM's capabilities on OpenEQA, a challenging benchmark for embodied video question answering. Our model achieves state-of-the-art performance, outperforming open-source models including strong Socratic models using GPT-4 as a planner by 3.6%. Additionally, we outperform Claude 3 and Gemini Pro Vision 1.0 and showcase competitive results compared to Gemini Pro 1.5 and GPT-4V, even surpassing the latter in spatial reasoning. This research paves the way for building efficient VLMs that can be deployed in robots or wearables, leveraging embodied video understanding to collaborate seamlessly with humans in everyday tasks, contributing to the next generation of Embodied AI.

For agents

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

get_harvested_code_for_paper("2406.13807")
get_code_for_paper("2406.13807")
have("2406.13807")

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

curl https://syntology.ai/api/ran/2406.13807.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.13807.svg)](https://syntology.ai/paper/2406.13807)

Connect an agent — have() is free.