United States

Diffusion-based models have been recently shown to be high-quality data generators. However, its performance severely degrades when training on non-stationary changing data distributions in an online manner, which is because of catastrophic forgetting. In this paper, we propose to enable the diffusion model with a novel Dynamic Expansion Memory Unit (DEMU) approach that adaptively creates new memory buffers to preserve the information deemed critical for the training of the model. Having a selective memory buffering unit is essential for training diffusion networks, which are expensive to train, especially when deployed in resource constrained environments. A Maximum Mean Discrepancy (MMD) based expansion mechanism, that evaluates probabilistic distances between each previously defined memory buffers and the newly given data, as an expansion signal, is employed for ensuring the learning of a diversity of information. A novel memory compaction approach is proposed to automatically remove statistically overlapping memory units through a graph relationship evaluation, is employed for preventing the limitless expansion of DEMU. In addition, we propose a new model expansion mechanism to automatically add new diffusion models as experts in a mixture system, which enhances the multi-domain image generation performance. Comprehensive results show that the proposed approach performs better than the state-of-the-art.

AAAI 2025

Lifelong Scalable Generative System via Online Maximum Mean Discrepancy

life-long and continual 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/)

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.



technical paper

Learning a perception and reasoning module for robotic assistants to plan steps to perform complex tasks based on natural language instructions often requires large free-form language annotations, especially for short high-level instructions. To reduce the cost of annotation, large language models (LLMs) are used as a planner with few data. However, when elaborating the steps, even the state-of-the-art planner that uses LLMs mostly relies on linguistic common sense, often neglecting the status of the environment at command reception, resulting in inappropriate plans. To generate plans grounded in the environment, we propose FLARE (Few-shot Language with environmental Adaptive Replanning Embodied agent), which improves task planning using both language command and environmental perception. As language instructions often contain ambiguities or incorrect expressions, we additionally propose to correct the mistakes using visual cues from the agent. The proposed scheme allows us to use a few language pairs thanks to the visual cues and significantly outperforms state-of-the-art approaches (by twice the success rate in unseen environments of the ALFRED benchmark: $16.42$\% $\to$ $40.88$\%).

Multi-Modal Grounded Planning and Efficient Replanning for Learning Embodied Agents with a Few Examples

Speaker recognition technology is applied to various tasks, from personal virtual assistants to secure access systems. However, the robustness of these systems against adversarial attacks, particularly to additive perturbations, remains a significant challenge. In this paper, we pioneer applying robustness certification techniques to speaker recognition, originally developed for the image domain. Our work covers this gap by transferring and improving randomized smoothing certification techniques against norm-bounded additive perturbations for classification and few-shot learning tasks to speaker recognition. We demonstrate the effectiveness of these methods on VoxCeleb 1 and 2 datasets for several models. We expect this work to improve the robustness of voice biometrics and accelerate the research of certification methods in the audio domain.

Certification of Speaker Recognition Models to Additive Perturbations

Large Language Models (LLMs) often face challenges when prompted to generate content under specific constraints or conditions. However, in such cases it is often easy to check whether these constraints are satisfied or violated. Recent works have shown that LLMs can benefit from such ``corrective feedback''. Here we claim that this skill of LLMs can be significantly enhanced via training. We introduce an RL framework for teaching models to use such rewards, by simulating interaction sessions, and rewarding the model according to its ability to satisfy the constraints.
We refer to our method as CORGI (Controlled Generation with RL for Guided Interaction), and evaluate it on a variety of controlled generation tasks using unlabeled training data. We find that CORGI consistently outperforms the baseline reinforcement learning method that does not incorporate conversational feedback. Furthermore, CORGI's interactive framework enables meta-learning, allowing the LLM to generalize better to guided interaction in new tasks. Through this dynamic interaction, the model learns to optimize its responses in real time, ensuring improved compliance with predefined requirements. Our experiments demonstrate that combining textual feedback with reinforcement learning results in more effective optimization and enhanced performance across a diverse range of generation tasks. These findings indicate that conversational optimization, when combined with reinforcement learning, significantly improves the effectiveness of LLMs in controlled generation contexts.

Teaching Models to Improve on Tape

Extracting lane topology from perspective views (PV) is crucial for planning and control in autonomous driving. This approach extracts potential drivable trajectories for self-driving vehicles without relying on high-definition (HD) maps. However, the unordered nature and weak long-range perception of the DETR-like framework can result in misaligned segment endpoints and limited topological prediction capabilities. Inspired by the learning of contextual relationships in language models, the connectivity relations in roads can be characterized as explicit topology sequences. In this paper, we introduce $\textbf{Topo2Seq}$, a novel approach for enhancing topology reasoning via topology sequences learning. The core concept of Topo2Seq is a randomized order prompt-to-sequence learning between lane segment decoder and topology sequence decoder. The dual-decoder branches simultaneously learn the lane topology sequences extracted from the Directed Acyclic Graph (DAG) and the lane graph containing geometric information. Randomized order prompt-to-sequence learning extracts unordered key points from the lane graph predicted by the lane segment decoder, which are then fed into the prompt design of the topology sequence decoder to reconstruct an ordered and complete lane graph. In this way, the lane segment decoder learns powerful long-range perception and accurate topological reasoning from the topology sequence decoder. Notably, topology sequence decoder is only introduced during training and does not affect the inference efficiency. Experimental evaluations on the OpenLane-V2 dataset demonstrate the state-of-the-art performance of Topo2Seq in topology reasoning.

Topo2Seq: Enhanced Topology Reasoning via Topology Sequence Learning

Today, deep neural networks are widely used since they can handle a variety of complex tasks. Their generality makes them very powerful tools in modern technology. However, deep neural networks are often overparameterized. The usage of these large models consumes a lot of computation resources. 

In this paper, we introduce a method called **T**ill the **L**ayers **C**ollapse (TLC), which compresses deep neural networks through the lenses of batch normalization layers. By reducing the depth of these networks, our method decreases deep neural networks' computational requirements and overall latency. We validate our method on popular models such as Swin-T, MobileNet-V2, and RoBERTa, across both image classification and natural language processing (NLP) tasks. The source code will be made publicly available upon the acceptance of this paper.

Till the Layers Collapse: Compressing a Deep Neural Network Through the Lenses of Batch Normalization Layers.

We study Consistent Query Answering (CQA) over knowledge bases with existential rules. Specifically, we propose a novel framework for CQA that combines previous approaches, allowing for the simultaneous presence of both open and closed predicates, i.e. predicates interpreted under open- and closed-world assumption, respectively. We establish the data complexity of answering unions of conjunctive queries in such a new framework under the so-called AR semantics and for different classes of existential rules. We also provide new complexity results for the standard (i.e. non-inconsistency tolerant) query answering in the presence of both open and closed predicates. Our results show that, for certain classes of rules, the complexity of CQA matches that of non-inconsistency-tolerant query answering.

Consistent Query Answering over Existential Rules with Open and Closed Predicates

Language-conditioned robotic manipulation in unstructured environments presents significant challenges for intelligent robotic systems. However, due to partial observation or imprecise action prediction, failure may be unavoidable for learned policies. Moreover, operational failures can lead to the robotic arm entering an untrained state, potentially causing destructive results. Consequently, the ability to detect and self-correct failures is crucial for the development of practical robotic systems. To address this challenge, we propose a foresight-driven failure detection and self-correction module for robot manipulation. By leveraging 3D Gaussian Splatting, we represent the current scene with multiple Gaussians. Subsequently, we train a prediction network to forecast the Gaussian representation of future scenes conditioned on planned actions. Failure is detected when the predicted future significantly deviates from the real observation after action execution. In such cases, the end-effector rolls back to the previous action to avoid an untrained state.  Integrating this approach with the PerACT framework, we develop a self-correcting robot manipulation policy. Evaluations on ten RLBench tasks with 166 variations demonstrate the superior performance of the proposed method, which outperforms state-of-the-art methods by 12.0%  success rate on average.

Self-Correcting Robot Manipulation via Gaussian-Splatted Foresight

Recent advancements in question generation (QG) have been significantly propelled by reinforcement learning (RL). Although extensive reward functions have been designed to capture the attributes of ideal questions, their associated learning challenges, particularly in sample efficiency and diversity, remain underexplored. This paper introduces a bilevel policy decomposition (BPD) framework and a diversity-seeking RL (DSRL) objective to address these issues. The BPD framework employs two cascading policies that divide QG into more manageable sub-tasks: answer-centric summary generation and summary-augmented QG, facilitating exploration and accelerating policy learning. Concurrently, the DSRL objective preserves the inherent diversity property of QG by ensuring the bilevel policies align probabilistically with their reward functions, rather than merely maximizing returns. Our integrated approach, termed BPD-DSRL, demonstrates superior performance over existing baselines on multiple question quality and diversity metrics across various QG benchmarks.

Enhancing Question Generation through Diversity-Seeking Reinforcement Learning with Bilevel Policy Decomposition

Dynamic interacting system modeling is important for understanding and simulating real world systems, e.g., meteorology and the spread of COVID. The system is typically described as a graph, where multiple objects dynamically interact with each other and evolve over time. In recent years, graph Ordinary Differential Equations (ODE) receive increasing research attentions. While achieving encouraging results, existing solutions prioritize the traditional Euclidean space, and neglect the intrinsic geometry of the system and physics laws, e.g., the principle of entropy increasing. The aforementioned limitations motivate us to rethink the system dynamics from a fresh perspective of Riemannian geometry, and pose a more realistic problem of physics-informed dynamic system modeling, considering the underlying geometry and physics law for the first time. In this paper, we present a novel physics-informed Riemannian graph ODE for a wide range of entropy-increasing dynamic systems (termed as Pioneer). In particular, we formulate a differential system on the Riemannian manifold, where a manifold-valued graph ODE is governed by the proposed constrained Ricci flow, and a manifold preserving Gyro-transform aware of system geometry. Theoretically, we report the provable entropy non-decreasing of our formulation, obeying the physics laws. Empirical results show the superiority of Pioneer on real datasets.

Pioneer: Physics-informed Riemannian Graph ODE for Entropy-increasing Dynamics

DNN-based watermarking methods have rapidly advanced, with the ``Encoder-Noise Layer-Decoder'' (END) framework being the most widely used. To ensure end-to-end training, the noise layer in the framework must be differentiable. However, real-world distortions are often non-differentiable, leading to challenges in end-to-end training. 
Existing solutions only treat the distortion perturbation as additive noise, which does not fully integrate the effect of distortion in training.
To better incorporate non-differentiable distortions into training, we propose a novel dual-decoder architecture ($\textbf{END$^{2}$}$). Unlike conventional END architecture, our method employs two structurally identical decoders: the Teacher Decoder, processing pure watermarked images, and the Student Decoder, handling distortion-perturbed images. 
The gradient is backpropagated only through the Teacher Decoder branch to optimize the encoder thus bypassing the problem of non-differentiability. To ensure resistance to arbitrary distortions, we enforce alignment of the two decoders' feature representations by maximizing the cosine similarity between their intermediate vectors on a hypersphere. 
Extensive experiments demonstrate that our scheme outperforms state-of-the-art algorithms under various non-differentiable distortions. Moreover, even without the differentiability constraint, our method surpasses baselines with a differentiable noise layer. Our approach is effective and easily implementable across all END architectures, enhancing practicality and generalizability.

Downloads

Next from AAAI 2025

Multi-Modal Grounded Planning and Efficient Replanning for Learning Embodied Agents with a Few Examples

Similar lecture

Continual Unsupervised Generative Modelling via Online Optimal Transport

.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 2025

Multi-Modal Grounded Planning and Efficient Replanning for Learning Embodied Agents with a Few Examples

Similar lecture

Continual Unsupervised Generative Modelling via Online Optimal Transport

Downloads