Survey of alignment techniques for large language models.
problem Ensuring large language models align with human values.
method Analysis of diverse alignment methods and training paradigms.
result Preference-based methods offer more flexibility for nuanced alignment.
Develops techniques to align word embeddings from different sources.
problem Aligning word embeddings from diverse datasets or methods.
method Simple closed-form techniques for optimal rotation, translation, and scaling.
result Maximizes cosine similarity and minimizes root mean squared errors.
The application of machine learning to bioinformatics problems is well established. Less well understood is the application of bioinformatics techniques to machine learning and, in particular, the representation of non-biological data as biosequences. The aim of this paper is to explore the effects of giving amino acid…
Novel methods robustify Gromov-Wasserstein distance for cross-domain alignment.
problem Robustifying Gromov-Wasserstein distance for cross-domain alignment.
method Three novel techniques derived from robust statistics to improve GW and its variants.
result Empirical validation shows superior resilience to contamination.
Paper presents a method to align unpaired samples across different modalities.
problem Challenges in collecting paired samples for multimodal representation learning.
method Uses propensity score alignment based on Rubin's framework to estimate a common space for unpaired samples.
result Optimal transport matching significantly improves alignment in real-world data.
Develops a topology-based test for AI model alignment and interpretability.
problem Testing and interpreting opaque AI models is difficult due to their complexity and lack of interpretability.
method Introduces a topology-based multi-modal alignment test to make AI models more interpretable.
result Demonstrates the effectiveness of the topology-based test in making AI model deployment and comparison more intuitive.
iMATCH aligns non-contiguous chunks for iSTS with ILP and outperforms other systems.
problem Addressing iSTS with explanatory chunk alignment and score assignment.
method iMATCH uses ILP for chunk alignment and Random Forest for similarity type/score assignment.
result iMATCH outperforms other systems in alignment score and overall score for students dataset.
This paper improves cross-domain learning using random forests for manifold alignment.
problem Improving cross-domain learning and feature integration.
method Semi-supervised manifold alignment using random forest proximities.
result Random forest proximities enhance downstream classification accuracy.
New framework transfers latent knowledge from weak to strong models.
problem Aligning superhuman LLMs with human feedback.
method Transfer learning framework using refinement approach.
result Proves weak-to-strong generalization is possible.
CARDS improves decoding efficiency and alignment quality for LLMs.
problem Efficiency bottlenecks in decoding-time alignment for LLMs.
method Cascade Reward Sampling (CARDS) with segment-level rejection sampling and uncertainty-based segmentation.
result Significant improvement in decoding efficiency and alignment quality.
MetFA aligns source and target domains for cross-device image classification.
problem Learning discriminative class boundaries across different domains.
method Distance metric guided feature alignment (MetFA) for domain-invariant and discriminative feature extraction.
result MetFA outperforms state-of-the-art methods in cross-device image classification.
A technique for clustering categorical data using ensembled dissimilarity matrices.
problem Clustering categorical data efficiently and accurately.
method Generate many dissimilarity matrices, average them, and extend to high dimensions using alignment techniques.
result Our method provides better clustering results, especially for genome sequences.
New method measures patient similarity over time, improving disease risk prediction.
problem Chronic diseases' varying progression rates and heterogeneous clinical presentations make patient comparison difficult.
method Subsequence alignment to account for pathophysiological misalignment and varying patient presentation times.
result Subsequence alignment outperforms global alignment in predicting disease progression.
Paper proposes a novel method for aligning hierarchical data using optimal transport in hyperbolic spaces.
problem Aligning hierarchical data like ontologies without external supervision.
method Optimal transport over hyperbolic spaces.
result The method outperforms standard embedding alignment techniques.
Unified framework simplifies DPO algorithms for LLM alignment.
problem Vast number of DPO variants complicates model alignment.
method Mutual information inspired unifying framework with flexible priors.
result Many DPO variants can be derived from the new framework.
The abstract explores connections between reinforcement learning, scaling, and diffusion.
problem Aligning reinforcement learning with human feedback and scaling techniques.
method Clarifying connections between reinforcement learning, scaling, and diffusion.
result Introducing a resampling approach for alignment and reward-directed diffusion models.
TTW aligns time-series faster and more accurately than existing methods.
problem Efficiently aligning multiple time-series signals with varying lengths.
method TTW uses a sinc convolutional kernel and gradient-based optimization for linear time and sequence complexity.
result TTW outperforms existing methods in time-series averaging and classification tasks.
PILAF optimizes reward models from human feedback for better policy alignment.
problem Creating accurate reward models from human feedback for policy optimization.
method Policy-Interpolated Learning for Aligned Feedback (PILAF) that explicitly aligns preference learning with maximizing underlying oracle reward.
result PILAF is optimal from both optimization and statistical perspectives, demonstrating strong performance in RLHF settings.
AOT aligns LLMs on distributional preferences via optimal transport.
problem Current LLM alignment techniques lack distributional level alignment.
method Alignment via Optimal Transport (AOT) aligns LLMs on unpaired preference data.
result AOT enables alignment by penalizing reward distribution violations.
Study benchmarks embedding-based entity alignment methods for KGs.
problem Align entities across different KGs using embeddings.
method Survey and categorize 23 embedding-based methods, propose new KG sampling algorithm, develop open-source library.
result Understood strengths and limitations of embedding-based methods.
Proposes a method for coarse graph alignment using sparse partial least squares.
problem Aligning graphs with community structures when there's no natural one-to-one mapping.
method Sparse partial least squares method incorporating observed graph structures and imposing sparsity.
result Demonstrates effectiveness in simulations.
Proposes a new method for manifold alignment using geometry-regularized twin autoencoders.
problem Traditional MA methods lack out-of-sample extension and real-world applicability.
method Guided representation learning with geometry-regularized twin autoencoders.
result Improves cross-domain generalization and robustness while maintaining alignment fidelity.
ReMixMatch improves semi-supervised learning with new techniques for data efficiency.
problem Improving semi-supervised learning with limited labeled data.
method Distribution alignment and augmentation anchoring with AutoAugment.
result Significantly more data-efficient, requiring less labeled data for similar accuracy.
Researchers prove it's impossible to partially recover graph alignments in certain conditions.
problem Recovering vertex correspondence between two random graphs with correlated edges.
method Used the probabilistic method to build automorphisms between tree components of a subcritical Erdös-Rényi graph.
result Proved an impossibility result for partial recovery in the sparse regime with constant average degree and correlation.
A new method for analyzing shapes using FDA techniques.
problem Statistical shape analysis of deformed contours.
method Functional Data Analysis (FDA) with basis expansion and principal component analysis.
result Successfully identifies deformation parameters and captures contour distributions.
The paper proposes a method to align AI models using conformal risk control.
problem Aligning AI models to meet end-user requirements in non-generative settings.
method Post-processing a pre-trained model to better align with a subset of functions using conformal risk control.
result A probabilistic guarantee that the resulting conformal interval around a model contains a function approximately satisfying a desired property.
Generative Adversarial Networks align manifold geometries instead of densities.
problem Aligning data manifolds between domains.
method Introduces MGM GAN with importance sampling and geometry penalty.
result Advantages of modeling manifold geometry over density.
End-to-end TTS framework uses hard alignment to improve accuracy.
problem End-to-end TTS systems struggle with accurate alignment between input text and output acoustic features.
method Proposes a constrained alignment scheme with hard monotonic alignments, marginalized during training.
result Improves alignment learning and prediction in end-to-end TTS systems.
A new method for aligning datasets without known correspondences.
problem Aligning datasets from different domains without labeled correspondences.
method Integrates MDS and Wasserstein Procrustes for joint optimization of embeddings and correspondences.
result Maps datasets to a common low-dimensional space without labeled correspondences.
New optimization technique for aligning points to lines, improving existing algorithms.
problem Minimizing distances between points and lines with given constraints.
method Combining techniques from computational geometry, combinatorics, and convex optimization.
result First constant-factor approximation algorithms for Points-to-Lines alignment with polynomial running time.
Proposes a new method for pattern localization in time series.
problem Locating a predefined sequence of patterns in a time series.
method Maps time series into a latent correlation space, then aligns them.
result Significant improvements over state-of-the-art methods.
New method improves domain adaptation by aligning sub-domains.
problem Real-world domain adaptation with shifted class distributions.
method Rebalanced sub-domain alignment and novel generalization bound.
result Improves performance in shifted class distribution scenarios.
Efficiently transfers style to content without distorting the content structure.
problem Arbitrary style transfer in computer vision.
method Rigid alignment of style features to content features.
result High-quality stylized images with intact content structure.
Paper tackles hidden game problem in AI alignment and language games.
problem Hidden game problem in AI alignment and language games.
method Developed a composition of regret minimization techniques to discover and exploit hidden structures.
result Achieved optimal external and swap regret bounds for rapid convergence to correlated equilibria.
Paper explores LSTM visualization techniques for ECGs.
problem Visualizing LSTM models for ECG classification.
method Four visualization techniques applied to ECGs, focusing on input deletion masks.
result Best technique is learning an input deletion mask to reduce class score.
Unified approach to domain generalization by aligning gradients and Hessians.
problem Developing models that generalize well across unseen domains.
method Moment Alignment, extending transfer measure to DG, aligning derivatives across domains.
result Moment Alignment unifies gradient and Hessian matching approaches, improving generalizability.
A new method LDHA improves fMRI data alignment for cognitive state validation.
problem Accurate functional alignment across different subjects for MVP analysis.
method LDHA incorporates LDA into CCA for supervised fMRI alignment.
result LDHA outperforms other HA methods in MVP analysis.
ELS framework improves safety alignment by dynamically steering LLMs towards helpful responses.
problem Over-Refusal in Aligned Large Language Models
method Fine-tuning free framework using an Energy-Based Model (EBM) to dynamically steer LLMs during inference.
result Extensive experiments show a significant reduction in false refusals (from 57.3% to 82.6%) while maintaining safety performance.
New research shows label refinement and weak training have limitations for aligning LLMs.
problem Limitations of refinement methods for aligning large language models.
method Analyzed probabilistic assumptions and alternative approaches to label refinement and weak training.
result Label refinement and weak training suffer from irreducible error, leaving a performance gap.
Improves text-dependent speaker verification using neural network supervectors and AUC optimization.
problem Enhance performance in text-dependent speaker verification systems.
method Proposes a supervector generation method and AUC optimization for neural networks.
result Improves system performance through novel alignment techniques and AUC optimization.
Improves text-to-image diffusion models using GFlowNets.
problem Aligning diffusion models with text descriptions.
method Post-training diffusion models with GFlowNets to generate high-reward images.
result Effective alignment of large-scale text-to-image diffusion models with reward information.
RePULSe improves language model alignment by reducing undesired outputs without sacrificing overall performance.
problem Aligning language models with human preferences while minimizing undesired outputs.
method Integrates probabilistic inference into RL training to reduce undesired outputs.
result RePULSe achieves a better balance between expected reward and undesired output probability.
Method evaluates disentanglement in DLVMs, including those not aligned with latent axes.
problem Evaluate disentanglement in DLVMs, especially those not aligned with latent axes.
method Proposes a statistical method to discover generative factors of a dataset.
result Empirically demonstrates the advantage of the method on two datasets.
New algorithm improves inference-time alignment without reward hacking.
problem Improving quality of responses from language models with limited compute.
method Inference-time alignment, focusing on extttInferenceTimePessimism algorithm. result Optimal performance and scaling-monotonicity of extttInferenceTimePessimism. The paper proposes a method to align surgical videos using kinematic data.
problem Difficulty in learning from comparing novice to expert surgical videos due to variability in gesture duration and execution.
method A novel technique using Dynamic Time Warping to synchronize videos of the same gesture at different speeds.
result The proposed approach allows for the alignment of surgical videos, enabling better learning for novice trainees.
Paper introduces MSA for weakly supervised covariance alignment in MEG signals.
problem Limited labeled signals in target datasets for MEG applications.
method Mixing model Stiefel Adaptation (MSA) leveraging unlabeled data.
result MSA outperforms recent methods in brain-age regression with MEG signals.
Hessian alignment improves OOD generalization in deep learning.
problem Improving deep learning models' ability to generalize to out-of-distribution data.
method Analyzed Hessian and gradient alignment for domain generalization using recent OOD theory.
result Hessian alignment methods achieve promising performance on various OOD benchmarks.
The paper develops dynamic word embeddings to capture evolving language structures.
problem Capturing the evolving meanings and associations of words over time.
method Develops a dynamic statistical model to learn time-aware word vector representation, solving the alignment problem.
result The model reliably captures the evolution of language over time and outperforms state-of-the-art approaches.