Model stitching compares neural representations, revealing insights not captured by CKA.
problem Understanding internal neural representations.
method Model stitching connects neural network layers to study representations.
result Good networks trained differently can be stitched without performance drop.
This paper challenges the conventional wisdom about temporal difference learning's superiority in stitching experience.
problem The effectiveness of temporal difference learning in stitching experience across different reinforcement learning methods.
method Empirical study comparing Monte Carlo and temporal difference learning methods in stitching experience.
result Monte Carlo methods can also achieve experience stitching, challenging the conventional wisdom.
Improves RL from historical data by stitching trajectories.
problem Lack of high-quality data for offline RL.
method Trajectory Stitching (TS) to augment historical data with synthetic actions.
result Improves RL policy performance over baseline.
Many real-world applications require robust algorithms to learn point processes based on a type of incomplete data --- the so-called short doubly-censored (SDC) event sequences. We study this critical problem of quantitative asynchronous event sequence analysis under the framework of Hawkes processes by leveraging the …
Recovering manifold geometry from geodesic intersections.
problem Recovering the geometry of a Riemannian manifold from geodesic intersection lengths.
method Applying stitching data to solve the delayed collision data problem.
result Geometry of the manifold can be recovered from geodesic intersection lengths.
A new approach selects tuning parameters for embedding methods.
problem Difficulty in selecting tuning parameters for embedding methods.
method Minimize a stress notion to supervise tuning parameter selection.
result Uncover a new bias--variance tradeoff phenomenon.
Detect spacetime curvature without rulers and clocks in 3D.
problem Detecting spacetime curvature without traditional measurement tools.
method Generalized results from 2D to 3D spacetime, proving well-stitched spacetime for conformally flat cases.
result A 3D spacetime is well-stitched if and only if it is conformally flat, providing a tool for curvature detection.
Task loss matching misrepresents similarity between neural network layers.
problem Measuring similarity between neural network layers using task loss matching.
method Task loss matching vs. direct matching; comparison with CCA and CKA.
result Direct matching provides a better similarity index than task loss matching.
The paper proposes a learning-theoretic perspective on representation alignment.
problem Understanding how AI models' representations become aligned as they scale.
method Reviewing and connecting different notions of alignment, focusing on stitching.
result Relating properties of stitching to kernel alignment of representations.
Improves BC policies by generating new plausible trajectories.
problem Sub-optimal data quality in BC leads to poor policy performance.
method Trajectory Stitching (TS) generates new plausible transitions.
result TS significantly improves behavioural policies over original data.
A new method for learning gradient flows from population dynamics.
problem Reconstructing population dynamics from limited data.
method Residual approach to enforce continuity equations, combining with data-fitting divergence.
result Demonstrated state-of-the-art performance across trajectory inference benchmarks.
A parametric point process model is developed, with modeling based on the assumption that sequential observations often share latent phenomena, while also possessing idiosyncratic effects. An alternating optimization method is proposed to learn a "registered" point process that accounts for shared structure, as well as…
This work presents a new robust PCA method for foreground-background separation on freely moving camera video with possible dense and sparse corruptions. Our proposed method registers the frames of the corrupted video and then encodes the varying perspective arising from camera motion as missing data in a global model.…
Research on manifold learning within a density ridge estimation framework has shown great potential in recent work for both estimation and de-noising of manifolds, building on the intuitive and well-defined notion of principal curves and surfaces. However, the problem of unwrapping or unfolding manifolds has received r…
In this paper, we identify an interesting kind of error in the output of Unsupervised Neural Machine Translation (UNMT) systems like \textit{Undreamt}(footnote). We refer to this error type as \textit{Scrambled Translation problem}. We observe that UNMT models which use \textit{word shuffle} noise (as in case of Undrea…
Proposes a continuous, differentiable model from local adaptive models.
problem Inadequate continuity and differentiability in over-parameterized models.
method A global continuous and differentiable model constructed from weighted averages of locally learned models.
result Achieves faster statistical convergence and improved performance in various settings.
Policy analysts wish to visualize a range of policies for large simulator-defined Markov Decision Processes (MDPs). One visualization approach is to invoke the simulator to generate on-policy trajectories and then visualize those trajectories. When the simulator is expensive, this is not practical, and some method is r…
Let M1 and M2 be two n-dimensional smooth manifolds with boundary. Suppose we glue M1 and M2 along some boundary components (which are, therefore, diffeomorphic). Call the result N. If we have a group G acting continuously on M1, and also acting continuously on M2, such that the actions are comp…
This paper generalizes graph representation for diverse data types.
problem Representing and querying hybrid data types in a unified way.
method Introducing a directed Tensor-Typed Multi-Graph with embeddings.
result Unified representation for visual, linguistic, and auditory data.
This work presents a novel approach for robust PCA with total variation regularization for foreground-background separation and denoising on noisy, moving camera video. Our proposed algorithm registers the raw (possibly corrupted) frames of a video and then jointly processes the registered frames to produce a decomposi…
The deep layers of modern neural networks extract a rather rich set of features as an input propagates through the network. This paper sets out to harvest these rich intermediate representations for quantization with minimal accuracy loss while significantly reducing the memory footprint and compute intensity of the DN…
Study links in knitted textiles using knot theory.
problem Identify and classify links in knitted textiles.
method Correspondence between links in thickened torus and 2-periodic weft-knitted textiles, using link invariants and ribbon knots.
result New stitch patterns and links in knitted textiles identified.
Classification is the most important process in data analysis. However, due to the inherent non-convex and non-smooth structure of the zero-one loss function of the classification model, various convex surrogate loss functions such as hinge loss, squared hinge loss, logistic loss, and exponential loss are introduced. T…
Extends FC-RAG to anytime-valid sequential coverage for language model swarms.
problem Maintain distribution-free coverage for a swarm of weak language models over time.
method Introduces Anytime-FC-RAG, a sequential extension with a summable calibration-deviation budget.
result Achieves time-uniform alarm validity and safety under predictable adaptive control.
This paper improves level generation using VAEs for coherent, logically following segments.
problem Generating coherent levels of non-fixed length and blending levels from different games.
method Sequential segment-based level generation using VAEs with a classifier for logical placement.
result Generated levels are more coherent and capable of blending levels from different games.
Linear-Core Surrogates combine fast optimization and statistical efficiency in classification and structured prediction.
problem The trade-off between smoothness and margin-based losses in classification and structured prediction.
method Linear-Core (LC) Surrogates, a family of convex loss functions that stitch a linear core to a smooth tail.
result LC Surrogates achieve fast linear consistency rates while maintaining differentiability and strict H-consistency bounds. Many signal processing algorithms break the target signal into overlapping segments (also called windows, or patches), process them separately, and then stitch them back into place to produce a unified output. At the overlaps, the final value of those samples that are estimated more than once needs to be decided in som…
CitySim dataset captures vehicle trajectories for safety research.
problem Lack of fine-grain vehicle trajectories for safety-oriented research.
method Five-step procedure: video stabilization, object filtering, stitching, detection, and error filtering.
result CitySim dataset improves safety evaluations and facilitates digital-twin research.
As Super-Resolution (SR) has matured as a research topic, it has been applied to additional topics beyond image reconstruction. In particular, combining classification or object detection tasks with a super-resolution preprocessing stage has yielded improvements in accuracy especially with objects that are small relati…
In several natural language tasks, labeled sequences are available in separate domains (say, languages), but the goal is to label sequences with mixed domain (such as code-switched text). Or, we may have available models for labeling whole passages (say, with sentiments), which we would like to exploit toward better po…
Bayesian RL enhances LLMs to reflectively explore and correct errors.
problem LLMs trained via RL lack reflective behaviors like rethinking and error correction.
method Bayesian RL framework that optimizes expected return under posterior distribution over Markov decision processes.
result BARL algorithm improves LLM performance in reasoning tasks.
Defines a bundle map for currents on manifolds using higher covariant derivatives.
problem Defining a bundle map for currents on manifolds.
method Using higher covariant derivatives on a manifold equipped with a torsion-free connection.
result The bundle of generalized Weyl algebras and its properties.
Improves local learning models for complex feature extraction.
problem Limited use of simple model families in local learning.
method Uses complex local model families to extract features.
result Demonstrates applications in various fields.
Local semi-supervised method improves brain tissue classification in child MRI.
problem Inaccurate detection of brain tissue classes due to intensity variations in early developing brains.
method Kernel Fisher Discriminant Analysis (KFDA) combined with SSIM for perceptual image quality assessment.
result Optimal brain partitioning into subdomains with different average intensity values and separating surfaces between brain parts.
This thesis tackles data imperfections in ML, proposing methods to prevent discrimination and spurious feature learning.
problem Data imperfections in ML training data, leading to deployment failures.
method Fair representation learning, spurious feature detection, and data augmentation for reinforcement learning.
result Methods to make ML models more robust to data imperfections.
This study compares mtl architectures for renewable power generation forecasting.
problem Forecasting expected power generation for renewable power plants.
method Comparative study of mtl architectures including Hard parameter sharing, cross-stitch network, sluice network, and task embedding.
result Emerging Relation Network (ERN) and task embedding provide significant improvements in forecasting.
Paper tackles identifying an odd arm in a multi-armed bandit with restless Markov processes and trembling hand.
problem Identifying an odd arm in a multi-armed bandit with restless Markov processes and trembling hand.
method Derive asymptotic lower bound on expected time to identify the odd arm, stitch together parameterised solutions to MDPs.
result First known asymptotic lower bound on expected time to identify the odd arm, with vanishing error probability.
A new method prioritizes and recycles experiences for better reinforcement learning.
problem Improving reinforcement learning efficiency by prioritizing and recycling experiences.
method Double-prioritized state-recycled (DPSR) experience replay.
result DPSR achieved state-of-the-art results in Atari games, outperforming original and prioritized methods.
This project combines recent advances in experience replay techniques, namely, Combined Experience Replay (CER), Prioritized Experience Replay (PER), and Hindsight Experience Replay (HER). We show the results of combinations of these techniques with DDPG and DQN methods. CER always adds the most recent experience to th…
ReaPER improves learning efficiency by prioritizing reliable experiences.
problem Inefficient sampling of past experiences in reinforcement learning.
method Introducing a novel measure of reliability to prioritize experiences in PER.
result ReaPER outperforms PER in various environments, including Atari-10.
Algorithm creates synthetic experiences to enhance Deep Reinforcement Learning.
problem Lack of synthetic experiences in classic Experience Replay.
method Bootstraps synthetic experiences to augment the replay memory.
result Synthetic experiences can improve learning speed and performance.
This paper describes an improvement in Deep Q-learning called Reverse Experience Replay (also RER) that solves the problem of sparse rewards and helps to deal with reward maximizing tasks by sampling transitions successively in reverse order. On tasks with enough experience for training and enough Experience Replay mem…
Unified model optimizes experiment performance and reduces duration.
problem Balancing reward maximization and experiment termination.
method Unified model that considers both within-experiment and post-experiment outcomes.
result Familiar algorithms can optimize a broad class of objectives with proper parameter adjustment.
New method designs experiments robustly for nonlinear estimation, improving parameter knowledge.
problem Designing robust experiments for nonlinear estimation under parametric uncertainty.
method Multi-stage robust optimization framework for sequential experiments.
result Identifies experiments better conducted early for improved parameter knowledge.
Experience reuse is key to sample-efficient reinforcement learning. One of the critical issues is how the experience is represented and stored. Previously, the experience can be stored in the forms of features, individual models, and the average model, each lying at a different granularity. However, new tasks may requi…
Optimal tests developed for sequential experiments with asymptotic properties.
problem Performing hypothesis tests after sequential experiments without prior design.
method Analyze asymptotic properties of sequential experiments; develop tests for Gaussian process observations.
result Asymptotic power function of any test can be matched by a specific test in a limit experiment.
Bayesian optimization for long-term outcomes using fast and slow experiments.
problem Optimizing long-term system effects with short-term misleading results.
method Combining fast and slow experiments for Bayesian optimization.
result Sequential optimization over large action spaces in a short time.
Two methods estimate effect size for online experiments, improving accuracy and efficiency.
problem Determining the correct effect size for online experiment duration.
method Two approaches: hierarchical models and utility theory.
result Proposed methods outperform baseline approaches in accuracy and efficiency.