12917 papers
This paper introduces AXIL, an exact, efficient, and scalable instance-attribution method for gradient boosting machines that computes prediction-specific weights via a matrix-free backward operator, outperforming existing methods in both faithfulness tests and computational speed while unifying exact decomposition with broader implicit differentiation frameworks.
This paper introduces Template-aware Dialogue Sentence Embedding (TaDSE), a novel self-supervised contrastive learning method that leverages easily obtainable token-level template information to generate high-quality sentence embeddings for task-oriented dialogues, achieving significant performance improvements over state-of-the-art methods on five benchmark datasets.
The paper introduces SciTune, a framework that aligns large language models with human-curated scientific multimodal instructions, resulting in a model (LLaMA-SciTune) that significantly outperforms state-of-the-art systems on scientific visual and language benchmarks, even surpassing human performance in certain categories.
The paper introduces MM-LIMA, a multimodal large language model that achieves superior performance compared to MiniGPT-4 by utilizing a trainable data selector to filter and fine-tune on a small, high-quality dataset of just 200 examples, demonstrating that less but higher-quality instruction data is more effective for alignment.
This paper proposes CROP, a model-based offline reinforcement learning algorithm that introduces a conservative reward estimator to mitigate distribution shift and overestimation by minimizing both estimation error and the rewards of random actions, achieving competitive performance through a streamlined objective.
This paper proposes PredFT, a novel fMRI-to-text decoding framework that leverages predictive coding theory by integrating brain predictive representations from specific regions of interest into a main network, thereby outperforming existing models in reconstructing continuous language from brain signals.
This paper proposes ITEM, an iterative framework inspired by Schutz's philosophical system of relevances that enhances the retrieval, utility judgment, and answer generation stages of Retrieval-Augmented Generation (RAG) systems, demonstrating significant improvements across multiple datasets.
This paper proposes LDNLM, a deep learning-based method that linearizes the traditional Nonlocal Means algorithm using a linear attention mechanism to achieve efficient and interpretable multiplicative noise removal with competitive performance.
This paper proposes a sample-efficient offline reinforcement learning approach for fixed-wing aircraft lateral attitude tracking that leverages system symmetry to augment training data and employs a dual-critic Deep Deterministic Policy Gradient (DDPG) structure to accelerate policy convergence.
This paper introduces "Deep Optimizer States," a novel technique that dynamically interleaves CPU and GPU computations by splitting model subgroups based on a performance model to exploit memory utilization fluctuations, thereby overcoming memory bottlenecks and achieving 2.5 faster training iterations compared to state-of-the-art offloading approaches.
This paper introduces Phantom, a framework that combines generative AI with a novel PCIe-specific constraint filter to eliminate hallucinations and synthesize high-fidelity, protocol-compliant Transaction Layer Packet (TLP) traces for practical device simulation.
The paper introduces WebLLM, an open-source JavaScript framework that leverages WebGPU and WebAssembly to enable high-performance, privacy-preserving large language model inference entirely within web browsers, achieving up to 80% of native device performance.
This paper introduces a privacy-preserving federated framework using threshold CKKS homomorphic encryption to enable multi-institutional Kaplan-Meier survival analysis that achieves high-fidelity results comparable to centralized data while preventing the reconstruction of sensitive individual records through encrypted aggregation and threshold decryption.
This paper demonstrates that human preference data quality for Reinforcement Learning from Human Feedback (RLHF) can be significantly improved by designing specific interventions—such as visualizing underlying model quantities, training users on the model, and modifying elicitation questions—to align human expression with the algorithm's preference model assumptions without altering their underlying reward functions.
This paper proposes AdaInit, a novel framework that leverages large language models with submartingale properties to iteratively synthesize adaptive initial parameters for Quantum Neural Networks, thereby effectively mitigating barren plateaus and ensuring higher gradient variance across various scales.
ExPath is a novel subgraph inference framework that integrates experimental molecular data and biological foundation models to identify biologically meaningful targeted pathways in knowledge bases, significantly outperforming existing explainers in fidelity and pathway length preservation.
This paper introduces LIFESTATE-BENCH, a novel benchmark utilizing narrative datasets like Hamlet to evaluate lifelong learning in large language models, revealing that while non-parametric methods outperform parametric ones in managing stateful interactions, all models still struggle with catastrophic forgetting over extended engagements.
This paper demonstrates that integrating strategic quantization and local inference techniques can reduce the energy consumption and carbon emissions of large language models by up to 45% without compromising their accuracy or responsiveness, offering a sustainable solution for resource-constrained environments.
This perspective paper explores the paradox where professional collaboration with AI accelerates the automation of expertise through knowledge externalization, while simultaneously proposing frameworks to transform and preserve human value through strategic adaptation in education, organizational design, and policy.
The paper introduces PODS, a method that accelerates reinforcement learning with verifiable rewards by decoupling rollout generation from policy updates and training only on a strategically down-sampled subset of rollouts, achieving the same performance as standard GRPO up to 1.7 times faster.