[GH Actions] Automatically add papers from authors (#78)

github-actions[bot] · RSTZZZ · web-flow · commit b832950d69d4 · 2025-06-15T16:20:56.000-04:00
Co-authored-by: RSTZZZ &lt;RSTZZZ@users.noreply.github.com&gt;
diff --git a/_posts/papers/2025-01-01-588454a41799f014e116b1d6718fb4571dff062e.md b/_posts/papers/2025-01-01-588454a41799f014e116b1d6718fb4571dff062e.md
@@ -0,0 +1,30 @@
+---
+title: 'BigDocs: An Open Dataset for Training Multimodal Models on Document and Code
+  Tasks'
+venue: International Conference on Learning Representations
+openAccessPdf:
+  url: ''
+  status:
+  license:
+names: Juan A. Rodriguez, Xiangru Jian, Siba Smarak Panigrahi, Tianyu Zhang, Aarash
+  Feizi, Abhay Puri, Akshay Kalkunte Suresh, François Savard, Ahmed Masry, Shravan
+  Nayak, Rabiul Awal, Mahsa Massoud, Amirhossein Abaskohi, Zichao Li, Suyuchen Wang,
+  Pierre-André Noël, Mats Leon Richter, Saverio Vadacchino, Shubham Agarwal, Sanket
+  Biswas, Sepideh Kharaghani
+tags:
+- International Conference on Learning Representations
+link: https://www.semanticscholar.org/paper/588454a41799f014e116b1d6718fb4571dff062e
+author: Aarash Feizi
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+None
diff --git a/_posts/papers/2025-02-21-2502.15210.md b/_posts/papers/2025-02-21-2502.15210.md
@@ -1,8 +1,16 @@
 ---
-title: 'PairBench: A Systematic Framework for Selecting Reliable Judge VLMs'
+title: 'PairBench: Are Vision-Language Models Reliable at Comparing What They See?'
 venue: ''
-names: Aarash Feizi, Sai Rajeswar, Adriana Romero-Soriano, Reihaneh Rabbany, Spandana
-  Gella, Valentina Zantedeschi, Joao Monteiro
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2502.15210, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Aarash Feizi, Sai Rajeswar, Adriana Romero-Soriano, Reihaneh Rabbany, Valentina
+  Zantedeschi, Spandana Gella, Joao Monteiro
 tags:
 - ''
 link: https://arxiv.org/abs/2502.15210
@@ -19,4 +27,4 @@ categories: Publications
 
 ## Abstract
 
-As large vision language models (VLMs) are increasingly used as automated evaluators, understanding their ability to effectively compare data pairs as instructed in the prompt becomes essential. To address this, we present PairBench, a low-cost framework that systematically evaluates VLMs as customizable similarity tools across various modalities and scenarios. Through PairBench, we introduce four metrics that represent key desiderata of similarity scores: alignment with human annotations, consistency for data pairs irrespective of their order, smoothness of similarity distributions, and controllability through prompting. Our analysis demonstrates that no model, whether closed- or open-source, is superior on all metrics; the optimal choice depends on an auto evaluator's desired behavior (e.g., a smooth vs. a sharp judge), highlighting risks of widespread adoption of VLMs as evaluators without thorough assessment. For instance, the majority of VLMs struggle with maintaining symmetric similarity scores regardless of order. Additionally, our results show that the performance of VLMs on the metrics in PairBench closely correlates with popular benchmarks, showcasing its predictive power in ranking models.
+Understanding how effectively large vision language models (VLMs) compare visual inputs is crucial across numerous applications, yet this fundamental capability remains insufficiently assessed. While VLMs are increasingly deployed for tasks requiring comparative judgment, including automated evaluation, re-ranking, and retrieval-augmented generation, no systematic framework exists to measure their performance in these scenarios. We present PairBench, a simple framework that evaluates VLMs as customizable similarity tools using widely available image datasets. Our approach introduces four key metrics for reliable comparison: alignment with human annotations, consistency across pair ordering, distribution smoothness, and controllability through prompting. Our analysis reveals that no model consistently excels across all metrics, with each demonstrating distinct strengths and weaknesses. Most concerning is the widespread inability of VLMs to maintain symmetric similarity scores. Interestingly, we demonstrate that performance on our benchmark strongly correlates with popular benchmarks used for more complex tasks, while providing additional metrics into controllability, smoothness and ordering. This makes PairBench a unique and comprehensive framework to evaluate the performance of VLMs for automatic evaluation depending on the task.
diff --git a/_posts/papers/2025-05-22-2505.16789.md b/_posts/papers/2025-05-22-2505.16789.md
@@ -0,0 +1,29 @@
+---
+title: 'Accidental Misalignment: Fine-Tuning Language Models Induces Unexpected Vulnerability'
+venue: ''
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2505.16789, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Punya Syon Pandey, Samuel Simko, Kellin Pelrine, Zhijing Jin
+tags:
+- ''
+link: https://arxiv.org/abs/2505.16789
+author: Kellin Pelrine
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+As large language models gain popularity, their vulnerability to adversarial attacks remains a primary concern. While fine-tuning models on domain-specific datasets is often employed to improve model performance, it can introduce vulnerabilities within the underlying model. In this work, we investigate Accidental Misalignment, unexpected vulnerabilities arising from characteristics of fine-tuning data. We begin by identifying potential correlation factors such as linguistic features, semantic similarity, and toxicity within our experimental datasets. We then evaluate the adversarial performance of these fine-tuned models and assess how dataset factors correlate with attack success rates. Lastly, we explore potential causal links, offering new insights into adversarial defense strategies and highlighting the crucial role of dataset design in preserving model alignment. Our code is available at https://github.com/psyonp/accidental_misalignment.
diff --git a/_posts/papers/2025-05-27-2505.20793.md b/_posts/papers/2025-05-27-2505.20793.md
@@ -0,0 +1,31 @@
+---
+title: Rendering-Aware Reinforcement Learning for Vector Graphics Generation
+venue: ''
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2505.20793, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Juan A. Rodriguez, Haotian Zhang, Abhay Puri, Aarash Feizi, Rishav Pramanik,
+  Pascal Wichmann, Arnab Mondal, Mohammad Reza Samsami, Rabiul Awal, Perouz Taslakian,
+  Spandana Gella, Sai Rajeswar, David Vázquez, Christopher Pal, Marco Pedersoli
+tags:
+- ''
+link: https://arxiv.org/abs/2505.20793
+author: Aarash Feizi
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+Scalable Vector Graphics (SVG) offer a powerful format for representing visual designs as interpretable code. Recent advances in vision-language models (VLMs) have enabled high-quality SVG generation by framing the problem as a code generation task and leveraging large-scale pretraining. VLMs are particularly suitable for this task as they capture both global semantics and fine-grained visual patterns, while transferring knowledge across vision, natural language, and code domains. However, existing VLM approaches often struggle to produce faithful and efficient SVGs because they never observe the rendered images during training. Although differentiable rendering for autoregressive SVG code generation remains unavailable, rendered outputs can still be compared to original inputs, enabling evaluative feedback suitable for reinforcement learning (RL). We introduce RLRF(Reinforcement Learning from Rendering Feedback), an RL method that enhances SVG generation in autoregressive VLMs by leveraging feedback from rendered SVG outputs. Given an input image, the model generates SVG roll-outs that are rendered and compared to the original image to compute a reward. This visual fidelity feedback guides the model toward producing more accurate, efficient, and semantically coherent SVGs. RLRF significantly outperforms supervised fine-tuning, addressing common failure modes and enabling precise, high-quality SVG generation with strong structural understanding and generalization.
diff --git a/_posts/papers/2025-06-01-2506.06347.md b/_posts/papers/2025-06-01-2506.06347.md
@@ -0,0 +1,30 @@
+---
+title: 'Unified Game Moderation: Soft-Prompting and LLM-Assisted Label Transfer for
+  Resource-Efficient Toxicity Detection'
+venue: ''
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2506.06347, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Zachary Yang, Domenico Tullo, Reihaneh Rabbany
+tags:
+- ''
+link: https://arxiv.org/abs/2506.06347
+author: Zachary Yang
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+Toxicity detection in gaming communities faces significant scaling challenges when expanding across multiple games and languages, particularly in real-time environments where computational efficiency is crucial. We present two key findings to address these challenges while building upon our previous work on ToxBuster, a BERT-based real-time toxicity detection system. First, we introduce a soft-prompting approach that enables a single model to effectively handle multiple games by incorporating game-context tokens, matching the performance of more complex methods like curriculum learning while offering superior scalability. Second, we develop an LLM-assisted label transfer framework using GPT-4o-mini to extend support to seven additional languages. Evaluations on real game chat data across French, German, Portuguese, and Russian achieve macro F1-scores ranging from 32.96% to 58.88%, with particularly strong performance in German, surpassing the English benchmark of 45.39%. In production, this unified approach significantly reduces computational resources and maintenance overhead compared to maintaining separate models for each game and language combination. At Ubisoft, this model successfully identifies an average of 50 players, per game, per day engaging in sanctionable behavior.
diff --git a/_posts/papers/2025-06-03-2506.02451.md b/_posts/papers/2025-06-03-2506.02451.md
@@ -0,0 +1,29 @@
+---
+title: Weak Supervision for Real World Graphs
+venue: ''
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2506.02451, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Pratheeksha Nair, Reihaneh Rabbany
+tags:
+- ''
+link: https://arxiv.org/abs/2506.02451
+author: Pratheeksha Nair
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+Node classification in real world graphs often suffers from label scarcity and noise, especially in high stakes domains like human trafficking detection and misinformation monitoring. While direct supervision is limited, such graphs frequently contain weak signals, noisy or indirect cues, that can still inform learning. We propose WSNET, a novel weakly supervised graph contrastive learning framework that leverages these weak signals to guide robust representation learning. WSNET integrates graph structure, node features, and multiple noisy supervision sources through a contrastive objective tailored for weakly labeled data. Across three real world datasets and synthetic benchmarks with controlled noise, WSNET consistently outperforms state of the art contrastive and noisy label learning methods by up to 15% in F1 score. Our results highlight the effectiveness of contrastive learning under weak supervision and the promise of exploiting imperfect labels in graph based settings.
diff --git a/_posts/papers/2025-06-03-2506.02873.md b/_posts/papers/2025-06-03-2506.02873.md
@@ -0,0 +1,31 @@
+---
+title: "It's the Thought that Counts: Evaluating the Attempts of Frontier LLMs to\
+  \ Persuade on Harmful Topics"
+venue: ''
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2506.02873, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Matthew Kowal, Jasper Timm, J. Godbout, Thomas H Costello, A. Arechar, Gordon
+  Pennycook, David Rand, Adam Gleave, Kellin Pelrine
+tags:
+- ''
+link: https://arxiv.org/abs/2506.02873
+author: Kellin Pelrine
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+Persuasion is a powerful capability of large language models (LLMs) that both enables beneficial applications (e.g. helping people quit smoking) and raises significant risks (e.g. large-scale, targeted political manipulation). Prior work has found models possess a significant and growing persuasive capability, measured by belief changes in simulated or real users. However, these benchmarks overlook a crucial risk factor: the propensity of a model to attempt to persuade in harmful contexts. Understanding whether a model will blindly ``follow orders'' to persuade on harmful topics (e.g. glorifying joining a terrorist group) is key to understanding the efficacy of safety guardrails. Moreover, understanding if and when a model will engage in persuasive behavior in pursuit of some goal is essential to understanding the risks from agentic AI systems. We propose the Attempt to Persuade Eval (APE) benchmark, that shifts the focus from persuasion success to persuasion attempts, operationalized as a model's willingness to generate content aimed at shaping beliefs or behavior. Our evaluation framework probes frontier LLMs using a multi-turn conversational setup between simulated persuader and persuadee agents. APE explores a diverse spectrum of topics including conspiracies, controversial issues, and non-controversially harmful content. We introduce an automated evaluator model to identify willingness to persuade and measure the frequency and context of persuasive attempts. We find that many open and closed-weight models are frequently willing to attempt persuasion on harmful topics and that jailbreaking can increase willingness to engage in such behavior. Our results highlight gaps in current safety guardrails and underscore the importance of evaluating willingness to persuade as a key dimension of LLM risk. APE is available at github.com/AlignmentResearch/AttemptPersuadeEval
diff --git a/_posts/papers/2025-06-03-2506.05393.md b/_posts/papers/2025-06-03-2506.05393.md
@@ -0,0 +1,30 @@
+---
+title: Are Large Language Models Good Temporal Graph Learners?
+venue: ''
+openAccessPdf:
+  url: ''
+  status:
+  license:
+  disclaimer: 'Notice: This content is from the open access paper or abstract available
+    at https://arxiv.org/abs/2506.05393, which is subject to the license by the author
+    or copyright owner provided with this content. Please go to the source to verify
+    the license and copyright information for your use.'
+names: Shenyang Huang, Alipanah Parviz, Emma Kondrup, Zachary Yang, Zifeng Ding, Michael
+  Bronstein, Reihaneh Rabbany, Guillaume Rabusseau
+tags:
+- ''
+link: https://arxiv.org/abs/2506.05393
+author: Shenyang Huang
+categories: Publications
+
+---
+
+*{{ page.names }}*
+
+**{{ page.venue }}**
+
+{% include display-publication-links.html pub=page %}
+
+## Abstract
+
+Large Language Models (LLMs) have recently driven significant advancements in Natural Language Processing and various other applications. While a broad range of literature has explored the graph-reasoning capabilities of LLMs, including their use of predictors on graphs, the application of LLMs to dynamic graphs -- real world evolving networks -- remains relatively unexplored. Recent work studies synthetic temporal graphs generated by random graph models, but applying LLMs to real-world temporal graphs remains an open question. To address this gap, we introduce Temporal Graph Talker (TGTalker), a novel temporal graph learning framework designed for LLMs. TGTalker utilizes the recency bias in temporal graphs to extract relevant structural information, converted to natural language for LLMs, while leveraging temporal neighbors as additional information for prediction. TGTalker demonstrates competitive link prediction capabilities compared to existing Temporal Graph Neural Network (TGNN) models. Across five real-world networks, TGTalker performs competitively with state-of-the-art temporal graph methods while consistently outperforming popular models such as TGN and HTGN. Furthermore, TGTalker generates textual explanations for each prediction, thus opening up exciting new directions in explainability and interpretability for temporal link prediction. The code is publicly available at https://github.com/shenyangHuang/TGTalker.
diff --git a/records/semantic_paper_ids_ignored.json b/records/semantic_paper_ids_ignored.json
@@ -12,6 +12,7 @@
   "05d76a00dd42fd5e75a7222d4479d2a35608c7a3",
   "05fac44a97728b033f2bbf32f273a8b20b5dce40",
   "06e81ae4a28d9c4c6e79e1ddff0c485a0a178ff1",
+  "07dc95fb09cfb8f76b9f28fb8ba639e0ee9dc109",
   "09a34aad92e6f416c81f47e60de0809616b49cce",
   "0b483b550b21ec42d693fc04a372dbb10dd07019",
   "0bb0ae024a30bbf67cfbff3775937136828f6c83",
@@ -83,6 +84,7 @@
   "37e06f3622c17dc6194b547c944462b2a513b878",
   "389596027e577fb28ea5e1cc313b4f3d610cffd3",
   "38ea195d707d1ed076052260036dbe5d7c1700c8",
+  "391cae6092b32ee1a9c8962aa161b8ea465dd11c",
   "3937b11717c22f62ab0b48dfd89e5dab75cedf40",
   "3a23eb4c2f2cebc8685bd9d01b3d7afed110b42a",
   "3af09b9c1dcc3c0177cc46c274be27596b3ab168",
@@ -116,13 +118,15 @@
   "511f44611a812b59184598767ea63c87afaddbe8",
   "5298c44e236fa58936cb40d55b11fa3f13dfa919",
   "5340b98af4349567189333209a95aedae988f6bc",
+  "53f74c8ba03079da8b05c5d895f5c9ff2badee21",
   "5411d9be13f73954715a56f1e25a686bcf296806",
   "544e342f3f263e49ad09d792f411410581ebd0f2",
   "54da615c0982af8c1285bd3d1ba9a9c00502a5a9",
   "55210ad2c71839d85653e2237e13403960c35b15",
   "56dc80163ef642b86ea2666bed33af97ef9d3b3b",
   "57133ef4c4de4d54a57686b8a914b06e4ff4aab5",
   "5751f36ea2fe917fc5536da17609aecd045be9ec",
+  "588454a41799f014e116b1d6718fb4571dff062e",
   "58d165b00a7cc20500e3b6b9fe02b603b9e97116",
   "59506544ad9b96c8731b9b78c3dd7b2b5114f67c",
   "59d2ce57f42f26d8bba11e5d76fa72a518eff00f",
@@ -193,6 +197,7 @@
   "898b14509593d235414df054527b7702e35c3099",
   "89cb62dc83c1b1895267bd28639fbf5bb7ed21a4",
   "8a0e8e3a495400220a34b3e0a78fe1534c1fbd5b",
+  "8a6abfef40ddc9ea571cd6fca37b66a852554b53",
   "8ad551a161b2af427dff42bed378828ec513aa7c",
   "8c12dae0297312dc3de5679d3c6ce34008007ab0",
   "8cc3d39f61d705331ecb4e66bda9b38e424c8f72",
@@ -241,6 +246,7 @@
   "ab5c71514254a1b2b3e2df31123dbc0257f4364d",
   "abd012e712c742e5325bd23abdb2ddf462308ddd",
   "ac2a659cc6f0e3635a9c1351c9963b47817205fb",
+  "ad28647fbd2624dab6eb6676bb9a40ff0f4238c2",
   "ad9232f48dc90540b28bbf3f3598dbc97d90b54f",
   "ada0dff98db47513a1659d078676c1671070c8ea",
   "adbbe35ca11f6f9bdb9f0f2643ea87fc6a083b6b",
@@ -322,8 +328,10 @@
   "e4273f01e6d7cd41357ea3af5c1a3df4962dc124",
   "e468a3b84707e9828bac47aef37815d5b4818a19",
   "e4b64ac277d3c4dc4567f3b5a15cb1fd17101944",
+  "e4c5f7cf65e3fa04cf1d4d68be9903244756bfdd",
   "e507aa05192ca5565d9f9ab2c50710aed01a7652",
   "e66b88e4dc9a6d5896748f9f40f7bbb5e67b0645",
+  "e67834207b4c2d2b170062e311115e05cc75982b",
   "e7e7445967768a168ba135e403477a25383f4d7c",
   "e874932b6a06171e35363dc9fe1d77936226028e",
   "e88fc995cf3acde90a0921e44f1890ba76360077",
