Skip to content

Commit

Permalink
Merge pull request #16 from aligebesce/main
Browse files Browse the repository at this point in the history
add mina, remove müge, add new paper
  • Loading branch information
gozdesahin committed Mar 8, 2024
2 parents 3cd3e0c + 7e178f0 commit 913be1c
Show file tree
Hide file tree
Showing 9 changed files with 70 additions and 31 deletions.
14 changes: 14 additions & 0 deletions _bibliography/papers.bib
Original file line number Diff line number Diff line change
@@ -1,5 +1,19 @@
---
---
@misc{uzunoglu2024paradise,
abbr = {arXiv},
bibtex_show = {true},
pdf = {2403.03167.pdf},
title={PARADISE: Evaluating Implicit Planning Skills of Language Models with Procedural Warnings and Tips Dataset},
author={Arda Uzunoglu and Abdalfatah Rashid Safa and Gözde Gül Şahin},
month={March},
year={2024},
url={https://arxiv.org/pdf/2403.03167.pdf},
eprint={2403.03167},
archivePrefix={arXiv},
primaryClass={cs.CL},
abstract = "Recently, there has been growing interest within the community regarding whether large language models are capable of planning or executing plans. However, most prior studies use LLMs to generate high-level plans for simplified scenarios lacking linguistic complexity and domain diversity, limiting analysis of their planning abilities. These setups constrain evaluation methods (e.g., predefined action space), architectural choices (e.g., only generative models), and overlook the linguistic nuances essential for realistic analysis. To tackle this, we present PARADISE, an abductive reasoning task using Q\&A format on practical procedural text sourced from wikiHow. It involves warning and tip inference tasks directly associated with goals, excluding intermediary steps, with the aim of testing the ability of the models to infer implicit knowledge of the plan solely from the given goal. Our experiments, utilizing fine-tuned language models and zero-shot prompting, reveal the effectiveness of task-specific small models over large language models in most scenarios. Despite advancements, all models fall short of human performance. Notably, our analysis uncovers intriguing insights, such as variations in model behavior with dropped keywords, struggles of BERT-family and GPT-4 with physical and abstract goals, and the proposed tasks offering valuable prior knowledge for other unseen procedural tasks. The PARADISE dataset and associated resources are publicly available for further research exploration with this https URL."
}

@misc{kural2024quantifying,
abbr = {arXiv},
Expand Down
9 changes: 4 additions & 5 deletions _members/phd_muge.md → _members/alumni_muge.md
Original file line number Diff line number Diff line change
@@ -1,17 +1,16 @@
---
layout: about
inline: false
group: PhD
group_rank: 2
team_frontpage: true
group: Former visitors, BSc/ MSc students, Interns
group_rank: 5
team_frontpage: false

title: MSc Müge Kural
description: Profile of Müge Kural, Doctoral Researcher at the GGLab.
description: Profile of Müge Kural, Ex Doctoral Researcher at the GGLab.
lastname: Kural
publications: 'author^=*Kural'

teaser: >
I am a second-year Ph.D. student at GGLab and KUIS AI.
I work on reasoning in LLMs.
My research interests include social reasoning in AI, human-AI interaction, and explainable AI.
Expand Down
25 changes: 25 additions & 0 deletions _members/intern_durhasan.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,25 @@
---
layout: about
inline: false
group: Undergrad Intern
group_rank: 4
team_frontpage: true

title: Mina Durhasan
description: Profile of Mina Durhasan, Bachelor Student at Koç University.
lastname: Durhasan
publications: 'author^=*Durhasan'

teaser: >
I am an undergrad researcher at GGLab majoring in CS at Koç University.
profile:
name: Mina Durhasan
align: right
image: mems/durhasan-profile.webp
role: Undergrad Intern
email: mdurhasan21@ku.edu.tr

---

I am an undergrad researcher at GGLab majoring in CS at Koç University.
25 changes: 0 additions & 25 deletions _members/mentee_ali.md

This file was deleted.

26 changes: 26 additions & 0 deletions _members/msc_ali.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,26 @@
---
layout: about
inline: false
group: Masters
group_rank: 3
team_frontpage: true

title: BSc Ali Gebeşçe
description: Profile of Ali Gebeşçe, Student Researcher at GGLab.
lastname: Gebeşçe
publications: 'author^=*Gebeşçe'

teaser: >
I earned my Bachelor of Science degree in Computer Science from Koç University and am a first-year M.Sc. student in Computer Science and Engineering at Koç University.
profile:
name: BSc Ali Gebeşçe
align: right
image: mems/gebesce-profile.webp
role: Undergrad Intern
email: agebesce17@ku.edu.tr

---

I earned my Bachelor of Science degree in Computer Science from Koç University and am a first-year M.Sc. student in Computer Science and Engineering at Koç University. Over the past year, I interned under the guidance of Asst. Prof. Gözde Gül Şahin at Koç University working at the topic of cognitive trust and the decision-making similarities between humans and NLP models.

File renamed without changes.
2 changes: 1 addition & 1 deletion _pages/publications.md
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@ layout: page
permalink: /publications
title: Publications
description: Publications by categories in reversed chronological order.
years: [2023]
years: [2024, 2023]
nav: true
nav_rank: 4
---
Expand Down
Binary file added assets/img/mems/durhasan-profile.webp
Binary file not shown.
Binary file added assets/pdf/2403.03167.pdf
Binary file not shown.

0 comments on commit 913be1c

Please sign in to comment.