Singapore

Bootstrapping large language models (LLMs) through preference-based policy optimization offers a promising direction for aligning model behavior with human preferences without relying on extensive manual annotations. In this work, we propose a novel online preference-based policy optimization (PbPO) framework that formulates the learning process as a min-max game between the main policy and a reward model (RM). The RM is constrained within a confidence set derived from preference data to ensure reliable exploitation. Our iterative online algorithm actively collects preference data through guided exploration of the evolving policy, enabling continual self-improvement of both the policy and the RM. We provide theoretical guarantees for our method, establishing a high-probability regret bound of order $\widetilde{\mathcal{O}}(d\sqrt{T})$, demonstrating its effectiveness in bootstrapping LLMs. Extensive experiments on five benchmarks show that our approach consistently outperforms existing state-of-the-art preference optimization techniques.

AAAI 2026

Bootstrapping LLMs via Preference-Based Policy Optimization

nlp: learning & optimization for nlp

nlp: (large) language models

ml: reinforcement learning

We are pleased to announce the Thirty-Ninth AAAI Conference on Artificial Intelligence (AAAI-25), which will be held in Philadelphia, Pennsylvania at the Pennsylvania Convention Center from February 25 to March 4, 2025.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-25 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.

### [Invited Speakers](https://aaai.org/conference/aaai/aaai-25/aaai-25-invited-speakers/)

Register [here](https://aaai.org/conference/aaai/aaai-25/registration/)

AAAI 2025

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-25 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.



poster

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-26 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to AAAI 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://aaai.org/conference/aaai/aaai-26/registration/) first.

**Online Registration Form**: https://aaai.getregistered.net/conference-2026 

Registration Required

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

Biological sequences, including RNAs and proteins, share similarities with natural languages, enabling the application of advanced language models to various biological tasks. However, due to its flexibility and lack of experimental data, RNA is a particularly challenging biological ``language'' compared to other biological sequences like proteins. RNA multiple sequence alignments (MSAs), which align evolutionarily related RNA sequences, can greatly enhance RNA biology modeling, as evidenced by their significant roles in structure prediction and function annotation. This raises the question of whether RNA MSAs can also benefit RNA design, which remains unexplored. This paper introduces RMSAGen, a model comprising RMSA-Encoder and RMSA-Decoder, that leverages MSAs to design functional RNA sequences. RMSA-Encoder effectively extracts MSA features, enhancing performance in functional prediction and solvent accessibility prediction tasks and supporting RMSA-Decoder in accurate RNA generation. RMSAGen can design RNA-binding protein sequences that effectively bind to target proteins, and the design performance improves with an increasing number of sequences. In addition, the ribozymes designed with structural features by RMSAGen show superior computational metrics and exhibit biological activity during gel electrophoresis. These results highlight effectiveness of RMSAGen, establishing it as a powerful tool and a new direction for RNA design.

RMSAGen: Integrating Multiple Sequence Alignment for Function RNA Design

Sharpness-aware minimization (SAM) is widely recognized for enhancing the generalization performance of deep neural networks. However, recent works have challenged the statement that flatness implies generalization, demonstrating that it is insufficient as the indicator of generalization \cite{Andriushchenko2023AML,Wen2023SharpnessMA}. In this paper, we reveal an insightful phenomenon: among minima of similar sharpness, stochastic optimization algorithms tend to prefer those with lower nonuniformity. We define nonuniformity by both the magnitude and structure of the gradient noise, and show that it fundamentally differs from sharpness and plays a critical role in generalization. Specifically, we first theoretically prove that the expected generalization gap of models trained via stochastic optimization algorithm is positively correlated with nonuniformity (the magnitude of the gradient noise). Empirically, we show that nonuniformity exhibits a stronger correlation with generalization than sharpness, especially in Transformer models. Furthermore, we demonstrate that the nonuniformity (the structure of the gradient noise) more effectively guides the algorithm towards sparser solutions and exhibits better generalization performance than sharpness-based methods in the high-dimensional sparse regression problem. Finally, extensive experiments on various datasets and models confirm the advantages of nonuniformity for generalization: (1) optimization guided by nonuniformity achieves better generalization compared to those achieved through flatness (including standard training, transfer learning, hyperparameter sensitivity and robustness to label noise); (2) model architecture (such as depth and width) is closely related to nonuniformity.

Beyond Sharpness: The Role of Nonuniformity in Generalization

As LLMs (large language models) are increasingly used to generate synthetic personas, particularly in data-limited domains such as health, privacy, and HCI, it becomes necessary to understand how these narratives represent identity, especially that of minority communities. In this paper, we audit synthetic personas generated by 3 LLMs (GPT4o, Gemini 1.5 Pro, Deepseek v2.5) through the lens of representational harm, focusing specifically on racial identity. Using a mixed-methods approach combining close reading, lexical analysis, and a parameterized creativity framework, we compare 1,512 LLM-generated persona to human-authored responses. Our findings reveal that LLMs disproportionately foreground racial markers, overproduce culturally coded language, and construct personas that are syntactically elaborate yet narratively reductive. These patterns result in a range of sociotechnical harms, including stereotyping, exoticism, erasure, and benevolent bias, that are often obfuscated by superficially positive narrations. We formalize this phenomenon as algorithmic othering, where minoritized identities are rendered hypervisible but less authentic. Based on these findings, we offer design recommendations for narrative-aware evaluation metrics and community-centered validation protocols for synthetic identity generation.

A Tale of Two Identities: An Ethical Audit of AI-Crafted Synthetic Personas

Large language models (LLMs) are increasingly deployed in real-world systems, making it critical to understand their vulnerabilities. While data poisoning attacks during RLHF/DPO alignment have been studied empirically, their theoretical foundations remain unclear. We investigate the minimum-cost poisoning attack required to steer an LLM’s policy toward an attacker’s target by flipping preference labels during RLHF/DPO, without altering the compared outputs. We formulate this as a convex optimization problem with linear constraints, deriving lower and upper bounds on the minimum attack cost. As a byproduct of this theoretical analysis, we show that any existing label-flipping attack can be post-processed via our proposed method to reduce the number of label flips required while preserving the intended poisoning effect. Empirical results demonstrate that this cost-minimization post-processing can significantly reduce poisoning costs over baselines, particularly when the reward model’s feature dimension is small relative to the dataset size. These findings highlight fundamental vulnerabilities in RLHF/DPO pipelines and provide tools to evaluate their robustness against low-cost poisoning attacks.

Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment

Backdoor attacks pose a serious threat to the security of large language models (LLMs), causing them to exhibit anomalous behavior under specific trigger conditions. The design of backdoor triggers has evolved from fixed triggers to dynamic or implicit triggers. This increased flexibility in trigger design makes it challenging for defenders to accurately identify their specific forms. Most existing backdoor defense methods are limited to specific types of triggers or rely on an additional clean model for support. To address this issue, we propose a backdoor detection method based on attention similarity, enabling backdoor detection without prior knowledge of the trigger. Our study reveals that models subjected to backdoor attacks exhibit unusually high similarity among attention heads when exposed to triggers. Based on this observation, we propose an attention safety alignment approach combined with head-wise fine-tuning to rectify potentially contaminated attention heads, thereby effectively mitigating the impact of backdoor attacks. Extensive experimental results demonstrate that our method significantly reduces the success rate of backdoor attacks while preserving the model’s performance on downstream tasks.

Uncovering and Aligning Anomalous Attention Heads to Defend Against NLP Backdoor Attacks

The deployment of decision-making AI agents presents a critical challenge in maintaining alignment with values or guidelines while operating in complex environments. Agents trained solely to achieve their task objectives may adopt harmful behaviors, exposing a key trade-off between maximizing reward and alignment. Avoiding misalignment is particularly difficult for pre-trained agents, where retraining is costly. This is further complicated by the diverse and potentially conflicting attributes representing ethical values. To address these challenges, we propose a test-time alignment technique based on model-guided policy shaping. Our method allows precise control over individual behavioral attributes, generalizes across diverse reinforcement learning (RL) environments, and facilitates a principled trade-off between ethical alignment and reward maximization without requiring agent retraining. We evaluate our approach using the MACHIAVELLI benchmark, which comprises 134 text-based game environments and thousands of annotated scenarios involving ethical decisions. The RL agents are first trained to maximize reward in their respective games. At test time, we apply policy shaping via scenario-action attribute classifiers to ensure decision alignment with ethical attributes. We compare our approach against prior training-time methods and general-purpose agents, and study several types of ethical violations and power-seeking behavior. Our results demonstrate that test-time policy shaping provides an effective and scalable solution for mitigating unethical behavior across diverse environments and alignment attributes.

Aligning Machiavellian Agents: Behavior Steering via Test-Time Policy Shaping

In the context of Visual Question Answering (VQA) and Agentic AI, calibration refers to how closely an AI system's confidence in its answers reflects their actual correctness. This aspect becomes especially important when such systems operate autonomously and must make decisions under visual uncertainty. While modern VQA systems, powered by advanced vision-language models (VLMs), are increasingly used in high-stakes domains like medical diagnostics and autonomous navigation due to their improved accuracy, the reliability of their confidence estimates remains under-examined. Particularly, these systems often produce overconfident responses. To address this, we introduce AlignVQA, a debate-based multi-agent framework, in which diverse specialized VLM -- each following distinct prompting strategies -- generate candidate answers and then engage in two-stage interaction: generalist agents critique, refine and aggregate these proposals. This debate process yields confidence estimates that more accurately reflect the model’s true predictive performance. We find that more calibrated specialized agents produce better aligned confidences. Furthermore, we introduce a novel differentiable calibration-aware loss function called AlignCal designed to fine-tune the specialized agents by minimizing an upper bound on the calibration error. This objective explicitly improves the fidelity of each agent’s confidence estimates. Empirical results across multiple benchmark VQA datasets substantiate the efficacy of our approach, demonstrating substantial reductions in calibration discrepancies.

Refine and Align: Confidence Calibration Through Multi-Agent Interaction in VQA

The rapid advancement of Large Language Model (LLM)-driven multi-agent systems has significantly streamlined software developing tasks, enabling users with little technical expertise to develop executable applications. While these systems democratize software creation through natural language requirements, they introduce significant security risks that remain largely unexplored. We identify two risky scenarios: Malicious User with Benign Agents (MU-BA) and Benign User with Malicious Agents (BU-MA). We introduce the Implicit Malicious Behavior Injection Attack (IMBIA), demonstrating how multi-agent systems can be manipulated to generate software with concealed malicious capabilities beneath seemingly benign applications, and propose Adv-IMBIA as a defense mechanism. Evaluations across ChatDev, MetaGPT, and AgentVerse frameworks reveal varying vulnerability patterns, with IMBIA achieving attack success rates of 93%, 45%, and 71% in MU-BA scenarios, and 71%, 84%, and 45% in BU-MA scenarios. Our defense mechanism reduced attack success rates significantly, particularly in the MU-BA scenario.
Further analysis reveals that compromised agents in the coding and testing phases pose significantly greater security risks, while also identifying critical agents that require protection against malicious user exploitation.
Our findings highlight the urgent need for robust security measures in multi-agent software development systems and provide practical guidelines for implementing targeted, resource-efficient defensive strategies.

Shadows in the Code: Exploring the Risks and Defenses of LLM-based Multi-Agent Software Development Systems

As large language models (LLMs) are increasingly deployed across culturally diverse regions, ensuring that their responses align with users’ cultural norms has become a critical challenge. Existing approaches to cultural alignment primarily rely on prompting or data-augmentation-based supervised finetuning, which teach models to follow norms indirectly through example-based supervision. However, these methods are difficult to scale and often fail to generalize, particularly in low-resource cultural settings.
In this work, we propose CultureRL, a culture-norm-driven reinforcement learning framework that directly encodes cultural principles into model behavior. Rather than relying on output imitation, CultureRL provides normative feedback during training, enabling the model to internalize high-level cultural rules. It consists of two key components: (1) Norm Pool Construction (NPC), which clusters data from the World Values Survey into abstract cultural concepts to form a structured and retrievable norm pool; and (2) Norm Cluster-based Reward Mechanism (NCRM), which retrieves the relevant norm for each input and uses an external reward model to assess conformity, guiding model updates through cultural alignment.
We evaluate CultureRL in both one-for-one (per-culture) and one-for-all (multi-culture) settings across nine cultures and three benchmarks. Results show that CultureRL consistently outperforms strong baselines, especially in terms of cultural consistency and adaptability.

CultureRL: Internalizing Cultural Principles in Large Language Models via Norm-Driven Reinforcement Learning

Scalable oversight protocols aim to empower evaluators to verify the output of AI models more capable than themselves. However, human evaluators are subject to biases that can lead to systematic errors. In a reanalysis of prior work that appeared to demonstrate the efficacy of simple protocols, we find that human evaluators possessing knowledge absent from models likely contributed to their positive results, which is an advantage that diminishes as models continue to scale in capability. We also report the results of two experiments examining the performance of simple oversight protocols where evaluators know that the model is "correct most of the time, but not all of the time'', finding no overall advantage for the tested protocols. In our main experiment, participants in both groups became more confident in the system’s answers after conducting online research, even when those answers were incorrect. These findings underscore the importance of testing the degree to which oversight protocols are robust to evaluator biases, whether they outperform a strategy of simple deference to the model being evaluated, and whether their performance scales with increasing problem difficulty and model capability.

Downloads

Next from AAAI 2026

RMSAGen: Integrating Multiple Sequence Alignment for Function RNA Design

Similar lecture

Forward KL Regularized Preference Optimization for Aligning Diffusion Policies

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from AAAI 2026

RMSAGen: Integrating Multiple Sequence Alignment for Function RNA Design

Similar lecture

Forward KL Regularized Preference Optimization for Aligning Diffusion Policies

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads