Finding the Sweet Spot: Preference Data Construction for Scaling Preference Optimization

Iterative data generation and model retraining are widely used to align large language models (LLMs). It typically involves a policy model to generate on-policy responses and a reward model to guide training data selection. Direct Preference Optimization (DPO) further enhances this process by constructing preference pairs of chosen and rejected responses. In this work, we aim to \emph{scale up} the number of on-policy samples via repeated random sampling to improve alignment performance. Conventional practice selects the sample with the highest reward as chosen and the lowest as rejected for DPO. However, our experiments reveal that this strategy leads to a \emph{decline} in performance as the sample size increases. To address this, we investigate preference data construction through the lens of underlying normal distribution of sample rewards. We categorize the reward space into seven representative points and systematically explore all 21 ($C_7^2$) pairwise combinations. Through evaluations on four models using AlpacaEval 2, we find that selecting the rejected response at reward position $\mu - 2\sigma$ rather than the minimum reward, is crucial for optimal performance. We finally introduce a scalable preference data construction strategy that consistently enhances model performance as the sample scale increases.

Sequence Transferability and Task Order Selection in Continual Learning

In continual learning, understanding the properties of task sequences and their relationships to model performance is important for developing advanced algorithms with better accuracy. However, efforts in this direction remain underdeveloped despite encouraging progress in methodology development. In this work, we investigate the impacts of sequence transferability on continual learning and propose two novel measures that capture the total transferability of a task sequence, either in the forward or backward direction. Based on the empirical properties of these measures, we then develop a new method for the task order selection problem in continual learning. Our method can be shown to offer a better performance than the conventional strategy of random task selection.

Knowledge-Guided Biomarker Identification for Label-Free Single-Cell RNA-Seq Data: A Reinforcement Learning Perspective

Gene panel selection aims to identify the most informative genomic biomarkers in label-free genomic datasets. Traditional approaches, which rely on domain expertise, embedded machine learning models, or heuristic-based iterative optimization, often introduce biases and inefficiencies, potentially obscuring critical biological signals. To address these challenges, we present an iterative gene panel selection strategy that harnesses ensemble knowledge from existing gene selection algorithms to establish preliminary boundaries or prior knowledge, which guide the initial search space. Subsequently, we incorporate reinforcement learning through a reward function shaped by expert behavior, enabling dynamic refinement and targeted selection of gene panels. This integration mitigates biases stemming from initial boundaries while capitalizing on RL's stochastic adaptability. Comprehensive comparative experiments, case studies, and downstream analyses demonstrate the effectiveness of our method, highlighting its improved precision and efficiency for label-free biomarker discovery. Our results underscore the potential of this approach to advance single-cell genomics data analysis.

Augmented Contrastive Clustering with Uncertainty-Aware Prototyping for Time Series Test Time Adaptation

Test-time adaptation aims to adapt pre-trained deep neural networks using solely online unlabelled test data during inference. Although TTA has shown promise in visual applications, its potential in time series contexts remains largely unexplored. Existing TTA methods, originally designed for visual tasks, may not effectively handle the complex temporal dynamics of real-world time series data, resulting in suboptimal adaptation performance. To address this gap, we propose Augmented Contrastive Clustering with Uncertainty-aware Prototyping (ACCUP), a straightforward yet effective TTA method for time series data. Initially, our approach employs augmentation ensemble on the time series data to capture diverse temporal information and variations, incorporating uncertainty-aware prototypes to distill essential characteristics. Additionally, we introduce an entropy comparison scheme to selectively acquire more confident predictions, enhancing the reliability of pseudo labels. Furthermore, we utilize augmented contrastive clustering to enhance feature discriminability and mitigate error accumulation from noisy pseudo labels, promoting cohesive clustering within the same class while facilitating clear separation between different classes. Extensive experiments conducted on three real-world time series datasets and an additional visual dataset demonstrate the effectiveness and generalization potential of the proposed method, advancing the underexplored realm of TTA for time series data.

Atomic Fact Decomposition Helps Attributed Question Answering

Attributed Question Answering (AQA) aims to provide both a trustworthy answer and a reliable attribution report for a given question. Retrieval is a widely adopted approach, including two general paradigms: Retrieval-Then-Read (RTR) and post-hoc retrieval. Recently, Large Language Models (LLMs) have shown remarkable proficiency, prompting growing interest in AQA among researchers. However, RTR-based AQA often suffers from irrelevant knowledge and rapidly changing information, even when LLMs are adopted, while post-hoc retrieval-based AQA struggles with comprehending long-form answers with complex logic, and precisely identifying the content needing revision and preserving the original intent. To tackle these problems, this paper proposes an Atomic fact decomposition-based Retrieval and Editing (ARE) framework, which decomposes the generated long-form answers into molecular clauses and atomic facts by the instruction-tuned LLMs. Notably, the instruction-tuned LLMs are fine-tuned using a well-constructed dataset, generated from large scale Knowledge Graphs (KGs). This process involves extracting one-hop neighbors from a given set of entities and transforming the result into coherent long-form text. Subsequently, ARE leverages a search engine to retrieve evidences related to atomic facts, inputting these evidences into an LLM-based verifier to determine whether the facts require expansion for re-retrieval or editing. Furthermore, the edited facts are backtracked into the original answer, with evidence aggregated based on the relationship between molecular clauses and atomic facts. Extensive evaluations demonstrate the superior performance of our proposed method over the state-of-the-arts on several datasets, with an additionally proposed new metric $Attr_{p}$ for evaluating the precision of evidence attribution.

Context-Aware Adapter Tuning for Few-Shot Relation Learning in Knowledge Graphs

Knowledge graphs (KGs) are instrumental in various real-world applications, yet they often suffer from incompleteness due to missing relations. To predict instances for novel relations with limited training examples, few-shot relation learning approaches have emerged, utilizing techniques such as meta-learning. However, the assumption is that novel relations in meta-testing and base relations in meta-training are independently and identically distributed, which may not hold in practice. To address the limitation, we propose RelAdapter, a context-aware adapter for few-shot relation learning in KGs designed to enhance the adaptation process in meta-learning. First, RelAdapter is equipped with a lightweight adapter module that facilitates relation-specific, tunable adaptation of meta-knowledge in a parameter-efficient manner. Second, RelAdapter is enriched with contextual information about the target relation, enabling enhanced adaptation to each distinct relation. Extensive experiments on three benchmark KGs validate the superiority of RelAdapter over state-of-the-art methods.

Diversified and Adaptive Negative Sampling on Knowledge Graphs

In knowledge graph embedding, aside from positive triplets (ie: facts in the knowledge graph), the negative triplets used for training also have a direct influence on the model performance. In reality, since knowledge graphs are sparse and incomplete, negative triplets often lack explicit labels, and thus they are often obtained from various sampling strategies (eg: randomly replacing an entity in a positive triplet). An ideal sampled negative triplet should be informative enough to help the model train better. However, existing methods often ignore diversity and adaptiveness in their sampling process, which harms the informativeness of negative triplets. As such, we propose a generative adversarial approach called Diversified and Adaptive Negative Sampling DANS on knowledge graphs. DANS is equipped with a two-way generator that generates more diverse negative triplets through two pathways, and an adaptive mechanism that produces more fine-grained examples by localizing the global generator for different entities and relations. On the one hand, the two-way generator increase the overall informativeness with more diverse negative examples; on the other hand, the adaptive mechanism increases the individual sample-wise informativeness with more fine-grained sampling. Finally, we evaluate the performance of DANS on three benchmark knowledge graphs to demonstrate its effectiveness through quantitative and qualitative experiments.

Double Oracle Neural Architecture Search for Game Theoretic Deep Learning Models

In this paper, we propose a new approach to train deep learning models using game theory concepts including Generative Adversarial Networks (GANs) and Adversarial Training (AT) where we deploy a double-oracle framework using best response oracles. GAN is essentially a two-player zero-sum game between the generator and the discriminator. The same concept can be applied to AT with attacker and classifier as players. Training these models is challenging as a pure Nash equilibrium may not exist and even finding the mixed Nash equilibrium is difficult as training algorithms for both GAN and AT have a large-scale strategy space. Extending our preliminary model DO-GAN, we propose the methods to apply the double oracle framework concept to Adversarial Neural Architecture Search (NAS for GAN) and Adversarial Training (NAS for AT) algorithms. We first generalize the players' strategies as the trained models of generator and discriminator from the best response oracles. We then compute the meta-strategies using a linear program. For scalability of the framework where multiple network models of best responses are stored in the memory, we prune the weakly-dominated players' strategies to keep the oracles from becoming intractable. Finally, we conduct experiments on MNIST, CIFAR-10 and TinyImageNet for DONAS-GAN. We also evaluate the robustness under FGSM and PGD attacks on CIFAR-10, SVHN and TinyImageNet for DONAS-AT. We show that all our variants have significant improvements in both subjective qualitative evaluation and quantitative metrics, compared with their respective base architectures.

Temporal Source Recovery for Time-Series Source-Free Unsupervised Domain Adaptation

Source-Free Unsupervised Domain Adaptation (SFUDA) has gained popularity for its ability to adapt pretrained models to target domains without accessing source domains, ensuring source data privacy. While SFUDA is well-developed in visual tasks, its application to Time-Series SFUDA (TS-SFUDA) remains limited due to the challenge of transferring crucial temporal dependencies across domains. Although a few researchers begin to explore this area, they rely on specific source domain designs, which are impractical as source data owners cannot be expected to follow particular pretraining protocols. To solve this, we propose Temporal Source Recovery (TemSR), a framework that transfers temporal dependencies for effective TS-SFUDA without requiring source-specific designs. TemSR features a recovery process that leverages masking, recovery, and optimization to generate a source-like distribution with recovered source temporal dependencies. To ensure effective recovery, we further design segment-based regularization to restore local dependencies and anchor-based recovery diversity maximization to enhance the diversity of the source-like distribution. The source-like distribution is then adapted to the target domain using traditional UDA techniques. Extensive experiments across multiple TS tasks demonstrate the effectiveness of TemSR, even surpassing existing TS-SFUDA method that requires source domain designs. Code is available in https://github.com/Frank-Wang-oss/TemSR.

EverAdapt: Continuous Adaptation for Dynamic Machine Fault Diagnosis Environments

Unsupervised Domain Adaptation (UDA) has emerged as a key solution in data-driven fault diagnosis, addressing domain shift where models underperform in changing environments. However, under the realm of continually changing environments, UDA tends to underperform on previously seen domains when adapting to new ones - a problem known as catastrophic forgetting. To address this limitation, we introduce the EverAdapt framework, specifically designed for continuous model adaptation in dynamic environments. Central to EverAdapt is a novel Continual Batch Normalization (CBN), which leverages source domain statistics as a reference point to standardize feature representations across domains. EverAdapt not only retains statistical information from previous domains but also adapts effectively to new scenarios. Complementing CBN, we design a class-conditional domain alignment module for effective integration of target domains, and a Sample-efficient Replay strategy to reinforce memory retention. Experiments on real-world datasets demonstrate EverAdapt superiority in maintaining robust fault diagnosis in dynamic environments. Our code is available: https://github.com/mohamedr002/EverAdapt