New framework improves LLM performance by avoiding forgetting during sequential training stages.
problem Forgetting during sequential training stages of LLMs.
method Proposes a joint post-training framework with theoretical convergence guarantees.
result Empirically outperforms sequential post-training framework by up to 23%.
Given a sequential learning algorithm and a target model, sequential machine teaching aims to find the shortest training sequence to drive the learning algorithm to the target model. We present the first principled way to find such shortest training sequences. Our key insight is to formulate sequential machine teaching…
A new method for Bayesian inference using diffusion models.
problem Bayesian inference in simulator-based models.
method Score-based diffusion models trained with a sequential training procedure.
result Comparable or superior performance compared to existing methods.
Language models trained on chess board states outperform those on moves, even with causal masking.
problem Applying causal masking to spatial data for training unimodal language models.
method Trained bidirectional and causal self-attention models on both spatial (board-based) and sequential (move-based) chess data.
result Models trained on spatial board states achieve stronger playing strength than those trained on sequential data, even with causal masking.
Proposes a new method for continual learning in neural networks.
problem Challenges in applying sequential Bayesian inference to neural networks.
method Sequential function-space variational inference.
result Neural networks trained with the proposed method achieve better predictive accuracy.
Boosted GFlowNets improve exploration by sequentially training GFlowNets with residual rewards.
problem GFlowNets struggle to evenly explore reward landscapes, leading to poor coverage of high-reward areas.
method Sequential training of an ensemble of GFlowNets, each optimizing a residual reward.
result Boosted GFlowNets achieve better exploration and sample diversity on multimodal benchmarks and peptide design tasks.
Dynamic SBI improves SBI efficiency without rounds, reducing simulation and training costs.
problem Efficiently perform complex scientific inference with high-dimensional data.
method Adaptive dataset transformation, parallel simulation and training.
result Significant improvements in simulation and training efficiency.
New algorithm for sequential off-policy learning improves performance over batch methods.
problem Training policies from logged interaction data in a sequential setting.
method Combines Logarithmic Smoothing with online PAC-Bayesian tools.
result Improves performance and accelerates convergence in sequential off-policy learning.
Study evaluates reinforcement learning algorithms for sequential experimental design.
problem Lack of generalization in reinforcement learning for experimental design.
method Investigated several reinforcement learning algorithms for sequential experimental design.
result Certain algorithms, using dropout or ensemble approaches, show attractive generalization properties.
New method uses randomized sparse neural networks to solve time-dependent PDEs more accurately and efficiently.
problem Numerical challenges in training neural networks sequentially in time to solve time-dependent PDEs.
method Introduces Neural Galerkin schemes that update randomized sparse subsets of network parameters at each time step.
result Up to two orders of magnitude more accurate and two orders of magnitude faster than dense update schemes.
Detects changes in classifier scores to identify shifts in class priors.
problem Label shift changes in classification data.
method Sequential changepoint detection of classifier scores.
result Outperforms other detection procedures in label shift settings.
Paper proposes a new method for sampling from complex distributions.
problem Sampling from unnormalised density functions in complex distributions.
method Combines amortised and particle-based methods with reinforcement learning.
result Improves sampling from complex distributions compared to existing methods.
We present Sequential Neural Likelihood (SNL), a new method for Bayesian inference in simulator models, where the likelihood is intractable but simulating data from the model is possible. SNL trains an autoregressive flow on simulated data in order to learn a model of the likelihood in the region of high posterior dens…
Proposed SMO algorithm for OC-SVM+ significantly outperforms non-sequential algorithms.
problem One-class SVM with privileged information
method Sequential Minimal Optimization (SMO) algorithm
result Finite-time convergence established
Survey on self-supervised pre-training for neural networks using unlabeled data.
problem Improving model performance using unlabeled data.
method Pre-training on unlabeled data followed by task-specific adaptation.
result Enhanced model performance through self-supervised pre-training.
A sequential learning framework improves domain generalization performance.
problem Training models robust to domain shift across multiple domains.
method Inspired by lifelong learning, a sequential training approach optimizing for all following domains.
result Improves performance on DG benchmarks with a simple, fast algorithm.
GP-TS optimizes TLM pre-training hyperparameters efficiently.
problem Resource inefficiency in TLM pre-training.
method Bayesian optimization with Thompson sampling and Gaussian process.
result GP-TS achieves lower MLM loss in fewer epochs.
Study shows how AI model can improve decision-making with missing data.
problem Sequential decision-making with missing covariates.
method Introduced model elasticity to quantify imputation discrepancy; used statistical learning and regression for calibration.
result Calibrating pre-trained models can significantly reduce regret in decision-making.
Improved diffusion models using energy distillation and sequential Monte Carlo.
problem Training instability and inferior performance in energy parameterized diffusion models.
method Introduced a novel training regime for energy functions through distillation of pre-trained diffusion models, and cast the sampling procedure as a Feynman Kac model.
result Demonstrated improved performance and new sampling techniques.
In this paper, we propose an AdaBoost-assisted extreme learning machine for efficient online sequential classification (AOS-ELM). In order to achieve better accuracy in online sequential learning scenarios, we utilize the cost-sensitive algorithm-AdaBoost, which diversifying the weak classifiers, and adding the forgett…
The paper addresses bias in fraud detection models by improving label recovery in payment networks.
problem Systematic bias in chargeback labels in payment networks.
method Formalizes the observation pipeline as a sequential missing-data problem with three stages and a corruption layer. Constructs the Sequential Triply Robust (STR) estimator to correct for all four impairments simultaneously.
result Achieves the semiparametric efficiency bound and provably dominates naive chargeback-based training in mean squared error.
Paper compares semi-supervised training for differentiable particle filters.
problem Lack of labelled data in real-world applications.
method Compares two semi-supervised training objectives.
result Improved performance in environments with scarce labelled data.
This paper speeds up OCSSVM training using SMO.
problem Training One-Class Slab SVMs is slow.
method Uses updated SMO to divide large problems into smaller, analytically solvable subproblems.
result Training OCSSVMs scales better with large datasets.
In the present work we propose a Deep Feed Forward network architecture which can be trained according to a sequential learning paradigm, where tasks of increasing difficulty are learned sequentially, yet avoiding catastrophic forgetting. The proposed architecture can re-use the features learned on previous tasks in a …
New reinforcement learning bound improves generalization for sequential data.
problem Challenges in obtaining generalization guarantees for reinforcement learning due to sequential data.
method PAC-Bayesian reinforcement learning with consideration of Markov dependencies and mixing time.
result Demonstrated practical utility through PB-SAC, providing meaningful confidence certificates.
New method selects features for sequential decision making.
problem Dynamic feature selection for instance-wise decisions.
method Latent variable model trained in a supervised manner; reasoning across stochastic latent space.
result Outperforms existing methods on various datasets.
Frequentist method estimates uncertainty in RNNs without altering architecture.
problem Uncertainty quantification in RNNs for decision-making.
method Jackknife resampling and influence functions to estimate variability.
result The method provides theoretical coverage guarantees on uncertainty intervals.
New SMC method for pBNNs improves scalability and predictive performance.
problem Training pBNNs with high-dimensional stochastic parameters.
method Gradient-based proposals within SMC samplers.
result New method outperforms state-of-the-art in predictive performance and training time.
New method improves ABI for sequential data, reducing forgetting and improving accuracy.
problem Performance degradation of ABI under model misspecification and distribution shifts.
method Decouples simulation-based pre-training from unsupervised SC fine-tuning, using memory buffer and elastic weight consolidation.
result Significant mitigation of forgetting and improved posterior estimates compared to standard simulation-based training.
In their standard form Gaussian processes (GPs) provide a powerful non-parametric framework for regression and classificaton tasks. Their one limiting property is their O(N3) scaling where N is the number of training data points. In this paper we present a framework for GP training with sequential sele…
We identify a phenomenon, which we refer to as multi-model forgetting, that occurs when sequentially training multiple deep networks with partially-shared parameters; the performance of previously-trained models degrades as one optimizes a subsequent one, due to the overwriting of shared parameters. To overcome this, w…
New method uses diffusion models for Bayesian inverse problems.
problem Solving Bayesian inverse problems with linear-Gaussian models.
method Decoupled Diffusion Sequential Monte Carlo (DDSMC) method.
result Asymptotically exact solution demonstrated on various data types.
SFSVI uses Gaussian mixtures to approximate neural network outputs for continual learning.
problem Learning new tasks without forgetting old ones in neural networks.
method Sequential function-space variational inference with Gaussian mixture approximation.
result Gaussian mixture SFSVI outperforms other methods in continual learning.
This paper presents a novel technique based on gradient boosting to train the final layers of a neural network (NN). Gradient boosting is an additive expansion algorithm in which a series of models are trained sequentially to approximate a given function. A neural network can also be seen as an additive expansion where…
Sequential hypothesis testing is a desirable decision making strategy in any time sensitive scenario. Compared with fixed sample-size testing, sequential testing is capable of achieving identical probability of error requirements using less samples in average. For a binary detection problem, it is well known that for k…
We study the problem of learning sequential decision-making policies in settings with multiple state-action representations. Such settings naturally arise in many domains, such as planning (e.g., multiple integer programming formulations) and various combinatorial optimization problems (e.g., those with both integer pr…
Recurrent neural networks (RNNs) are powerful and effective for processing sequential data. However, RNNs are usually considered "black box" models whose internal structure and learned parameters are not interpretable. In this paper, we propose an interpretable RNN based on the sequential iterative soft-thresholding al…
Current training regimes for deep learning usually involve exposure to a single task / dataset at a time. Here we start from the observation that in this context the trained model is not given any knowledge of anything outside its (single-task) training distribution, and has thus no way to learn parameters (i.e., featu…
Symmetric CNNs improve sequential recommendation and protein structure prediction.
problem Improving prediction accuracy in sequential recommendation and protein structure inference.
method Developed a CNN architecture that preserves symmetry in convolutional layers, using parameterized convolutional kernels.
result Symmetric structured CNNs achieve better performance with fewer parameters.
Improved latent dynamics identification framework reduces training time and improves accuracy.
problem Accurate numerical solutions of partial differential equations require computationally expensive solvers.
method Sequential decoder training (mLaSDI) to correct residual errors from previous stages.
result mLaSDI consistently outperforms standard LaSDI, achieving lower prediction errors and reduced training time.
Paper develops an efficient approach to reduce HPO time.
problem Challenges in determining optimal hyperparameters due to large number and training time.
method Nested Latin hypercube design for initialization, truncated additive Gaussian process model for calibration, sequential model-based algorithm for optimization.
result Demonstrates competitive performance on various machine learning models.
We build on auto-encoding sequential Monte Carlo (AESMC): a method for model and proposal learning based on maximizing the lower bound to the log marginal likelihood in a broad family of structured probabilistic models. Our approach relies on the efficiency of sequential Monte Carlo (SMC) for performing inference in st…
GIB improves neural network generalization by dynamically selecting task-relevant features across different sequential environments.
problem Poor generalization of deep neural networks to unseen environments.
method Proposes a gated information bottleneck (GIB) approach that dynamically drops spurious correlations and selects task-relevant features.
result GIB outperforms other IB approaches in adversarial robustness and OOD detection.
Constructs classifiers for neural networks with specific data configurations.
problem Finding global minima of deep ReLU neural networks on sequentially separable data.
method Explicitly constructs zero loss neural network classifiers using cumulative parameters and truncation maps.
result Global minimizers can be described with a limited number of parameters based on the data structure.
Discrete Flow Maps bypass sequential prediction limits for parallel text generation.
problem Sequential autoregressive prediction limits large language model speed.
method Flow Maps compress generative trajectories into single-step mappings.
result Discrete Flow Maps surpass previous state-of-the-art results in discrete flow modeling.
Sequential learning of multiple tasks in artificial neural networks using gradient descent leads to catastrophic forgetting, whereby previously learned knowledge is erased during learning of new, disjoint knowledge. Here, we propose a new approach to sequential learning which leverages the recent discovery of adversari…
New method improves training-free guidance for diffusion models, achieving state-of-the-art results.
problem Accurate, training-free guidance for conditional generation in diffusion models.
method Sequential Monte Carlo (SMC) framework with Multi-Level Monte Carlo (MLMC) variance reduction.
result Achieves state-of-the-art results on CIFAR-10 and ImageNet datasets with significant cost reduction.
Context-aware recommender systems (CARSs) apply sensing and analysis of user context in order to provide personalized services. Adding context to a recommendation model is challenging, since the addition of context may increases both the dimensionality and sparsity of the model. Recent research has shown that modeling …