Proposes FIPO-BC for efficient online calibration of complex models.
problem Efficiently calibrating computationally expensive models with large datasets.
method Fixed inducing points online Bayesian calibration (FIPO-BC) algorithm.
result FIPO-BC is at least ten times faster than standard methods and enables online updates.
BC learning improves deep sound recognition performance.
problem Improving deep sound recognition using novel training data.
method BC learning: mixing sounds from different classes to generate between-class sounds and train models to recognize these.
result BC learning improves performance on various sound recognition networks, surpassing human level.
The minimum description length (MDL) principle in supervised learning is studied. One of the most important theories for the MDL principle is Barron and Cover's theory (BC theory), which gives a mathematical justification of the MDL principle. The original BC theory, however, can be applied to supervised learning only …
Traditional GANs use a deterministic generator function (typically a neural network) to transform a random noise input z to a sample x that the discriminator seeks to distinguish. We propose a new GAN called Bayesian Conditional Generative Adversarial Networks (BC-GANs) that use a random generator function…
Improved RL policies from offline data with relaxed BC constraints.
problem Overestimation bias in offline RL due to lack of interaction with environment.
method Introducing a policy constraint via behavioural cloning (BC) and adjusting the balance between RL and BC.
result Refined policies outperform baseline and match/exceed complex alternatives.
New method learns robot skills from data, matching or outperforming existing methods.
problem Learning robot skills from fixed datasets.
method Offline Reinforcement Learning via Supervised Learning using implicit models.
result Implicit models can match or outperform explicit models in acquiring robotic skills.
This paper develops a novel graph neural network to efficiently identify high betweenness centrality nodes.
problem Efficiently identifying high betweenness centrality nodes in large networks.
method A novel encoder-decoder framework using pairwise ranking loss.
result The model accurately identifies highly-ranked nodes without noticeable sacrifice in accuracy.
Interactive IL beats BC by state-wise annotation cost.
problem Behavior Cloning struggles with annotation cost in sequential decision making.
method Proved Stagger and Warm Stagger algorithms to outperform BC.
result Interactive and hybrid IL methods outperform BC with state-wise annotation.
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.
BC learning improves image classification by mixing images from different classes.
problem Improving image classification accuracy.
method Generates mixed images from different classes and trains models to output the mixing ratio.
result Significant improvement in image classification performance (19.4% and 2.26% top-1 errors on ImageNet-1K and CIFAR-10, respectively).
Paper tackles offline RL from mixed datasets with adaptive KL regularizer.
problem Challenges in optimizing RL and BC signals with varying action coverage and multiple action modes.
method Adaptively weighted reverse KL divergence regularizer based on TD3 algorithm.
result Empirically outperforms existing offline RL algorithms in MuJoCo locomotion tasks.
COMBO network improves optical flow estimation by combining deep learning with brightness constancy.
problem Optical flow estimation using deep learning requires complex training schemes.
method COMBO network explicitly exploits brightness constancy and combines it with a data-driven approach.
result COMBO network outperforms state-of-the-art methods on various benchmarks.
Let $Q^N_l\subset \bC\bP^{N+1}$ denote the standard real, nondegenerate hyperquadric of signature l and $M\subset \bC^{n+1}$ a real, Levi nondegenerate hypersurface of the same signature l. We shall assume that there is a holomorphic mapping $H_0\colon U\to \bC\bP^{N_0+1}$, where U is some neighborhood of M in …
ORIL learns a reward function from unlabeled data to improve robot learning.
problem Leveraging unlabeled data for robot learning.
method ORIL learns a reward function from demonstrator and unlabeled trajectories, annotates data, and trains an agent via offline reinforcement learning.
result ORIL consistently outperforms BC agents on various robotic tasks.
We consider local CR-immersions of a strictly pseudoconvex real hypersurface $M\subset\bC^{n+1}$, near a point p∈M, into the unit sphere $\mathbb S\subset\bC^{n+d+1}$ with d>0. Our main result is that if there is such an immersion f:(M,p)→S and d<n/2, then f is {\em rigid} in the sense t…
The BC(n) Sutherland Hamiltonian with coupling constants parametrized by three arbitrary integers is derived by reductions of the Laplace operator of the group U(N). The reductions are obtained by applying the Laplace operator on spaces of certain vector valued functions equivariant under suitable symmetric subgroups o…
BinaryConnect is generalized and proven to converge.
problem Understanding and improving neural network quantization.
method Refined analysis of BC, introduction of ProxConnect.
result ProxConnect achieves competitive performance.
This paper uses Gaussian processes to cluster BC coastal rainfall patterns.
problem How to cluster BC coastal rainfall patterns effectively.
method Developed an approach for clustering multiple Gaussian processes observed on a comparable interval.
result Interesting insights into BC rainfall patterns, not simply clustering El Niño and La Niña years.
Study on harmonic forms on almost Hermitian 4-manifolds, calculating dimensions and invariants.
problem Understanding harmonic forms on almost Hermitian 4-manifolds.
method Analyzing Bott-Chern and ∂ˉ harmonic forms, calculating dimensions and invariants. result Dimensions of harmonic forms on almost Hermitian 4-manifolds are determined.
New symplectic embedding obstructions found for polydisks into half-integer ellipsoids.
problem Obstructing symplectic embeddings of polydisks into half-integer ellipsoids.
method Combinatorial criterion developed by Hutchings to obstruct symplectic embeddings.
result Optimal inclusion conditions for symplectic embeddings of polydisks into half-integer ellipsoids.
Study calculates global sections on special geometric spaces.
problem Calculating global sections on compact Ricci-flat Kähler manifolds.
method Expressed as an invariant subspace of a βγ-bc system under Lie algebra action.
result Space of global sections is an invariant subspace.
BC-ACI corrects time series forecast bias, improving prediction intervals.
problem Persistent bias in time series forecasts leads to overly conservative prediction intervals.
method Augments ACI with an EWM estimate of forecast bias to correct nonconformity scores and re-center intervals.
result Reduces Winkler interval scores by 13-17% under distribution shifts, improving calibration.
BC-LLM uses LLMs to find concepts without predefined sets, improving interpretability and performance.
problem Finding a balance between interpretability and accuracy in concept extraction models.
method Bayesian approach with LLMs as both concept extractor and prior.
result BC-LLM outperforms interpretable and black-box models across various datasets.
In this paper, we look to address the problem of estimating the dynamic direction of arrival (DOA) of a narrowband signal impinging on a sensor array from the far field. The initial estimate is made using a Bayesian compressive sensing (BCS) framework and then tracked using a Bayesian compressed sensing Kalman filter (…
Unified framework for policy learning using weak supervision.
problem High-quality supervision is often infeasible or expensive in practice.
method Treat weak supervision as imperfect peer information and evaluate policies based on correlated agreement.
result Substantial performance improvements, especially in complex or noisy environments.
The paper extends Bott-Chern Laplacian definition and explores its properties on almost Hermitian manifolds.
problem Exploring the properties of Bott-Chern Laplacian on almost Hermitian manifolds.
method Extending the definition of Bott-Chern Laplacian, proving ellipticity, and analyzing kernels on different types of manifolds.
result The dimensions of Bott-Chern and Dolbeault harmonic forms differ on almost complex 4-manifolds with specific metrics.
Unified probabilistic perspective on imitation learning methods using divergence minimization.
problem Understanding and improving imitation learning methods for limited demonstration scenarios.
method Unified probabilistic perspective based on divergence minimization.
result State-marginal matching objective contributes most to IRL's superior performance.
SQIL uses a simple reward strategy to encourage long-horizon imitation of expert demonstrations.
problem Challenges in imitation learning with high-dimensional, continuous observations and unknown dynamics.
method Imitates expert demonstrations by providing a constant reward of +1 for matching actions in demonstrated states, and 0 for all others.
result Empirically outperforms behavioral cloning and achieves competitive results compared to GAIL.
A new neural network architecture reduces parameters and improves performance.
problem Reducing model complexity and improving performance in neural networks.
method Coupled ensembles of neural networks with parallel branches and tighter coupling.
result Significant performance improvements with reduced parameter count.
In this paper, we consider real hypersurfaces M in C3 (or more generally, 5-dimensional CR manifolds of hypersurface type) at uniformly Levi degenerate points, i.e. Levi degenerate points such that the rank of the Levi form is constant in a neighborhood. We also require the hypersurface to satisfy a certain s…
Combines BC and GAIL for efficient imitation learning.
problem Efficient imitation learning without reward signals.
method Integrates Behavior Cloning and Generative Adversarial Imitation Learning.
result Combination leads to stable and sample-efficient learning.
Registration, which aims to find an optimal 1-1 correspondence between shapes, is an important process in different research areas. Conformal mappings have been widely used to obtain a diffeomorphism between shapes that minimizes angular distortion. Conformal registrations are beneficial since it preserves the local ge…
Researchers recover Riemannian manifolds and lower order terms from travel time data.
problem Recovering Riemannian manifolds and lower order terms from travel time data.
method Adaptation of the Boundary Control method to recover lower order terms.
result Complete Riemannian manifolds and lower order terms can be uniquely recovered from a local source to solution map.
A topological invariant of a polynomial map p:X→B from a complex surface containing a curve C⊂X to a one-dimensional base is given by a rational second homology class in the compactification of the moduli space of genus g curves with n labeled points $\modmgn$. Here the generic fibre of p has genus …
Decomposes harmonic forms on almost Kähler manifolds, revealing non-trivial structure.
problem Primitive decomposition of harmonic forms on compact almost Kähler manifolds.
method Primitive decomposition of ∂ˉ,∂, Bott-Chern and Aeppli-harmonic (k,k)-forms. result Primitive components of harmonic forms are constants multiples of ωk. This paper describes how to recover the topology of a closed manifold M from a good Morse function f on M. The essential method was suggested by Cohen, Jones and Segal. They constructed a topological category Cf and claimed that the classifying space BCf is homeomorphic to M. We prove it from a differ…
BC-Aligner maintains backward compatibility of embeddings after frequent updates.
problem Updating embeddings without requiring consumer teams to retrain their models.
method Learning backward compatible embeddings through BC-Aligner.
result BC-Aligner maintains backward compatibility with existing unintended tasks after multiple model version updates.
Causal Imitation Learning handles noisy measurements and distribution shifts.
problem Learning from noisy state observations and distributional shifts.
method Causal inference framework and adversarial RKHS learning.
result Improved robustness to distribution shifts compared to standard methods.
In this paper, we consider asymptotic properties of the support vector machine (SVM) in high-dimension, low-sample-size (HDLSS) settings. We show that the hard-margin linear SVM holds a consistency property in which misclassification rates tend to zero as the dimension goes to infinity under certain severe conditions. …
New polynomial connects knot genus to 3-manifold geometry.
problem Detecting knot genus from 3-manifold geometry.
method Ideal triangulation and super-Ptolemy assignments.
result New polynomial conjecturally agrees with torsion polynomial.
The paper shows conditions under which certain cohomology groups vanish for Hermitian manifolds.
problem Conditions for vanishing certain cohomology groups on Hermitian manifolds.
method Analyzes the cohomology groups of Hermitian manifolds with specific conditions on metrics and forms.
result If the Aeppli class of ωn−p vanishes, then HBCp,0(M)=0. The paper optimizes UAV path and power for QoS in cellular networks.
problem Optimizing UAV path and power for QoS in cellular networks.
method Apprenticeship learning via deep inverse reinforcement learning (IRL) combined with Q-learning and DRL.
result The proposed method achieves expert-level performance and maintains performance in unseen situations.
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.
Compressed sensing is a powerful tool in applications such as magnetic resonance imaging (MRI). It enables accurate recovery of images from highly undersampled measurements by exploiting the sparsity of the images or image patches in a transform domain or dictionary. In this work, we focus on blind compressed sensing (…
We propose a method for explicit computation of the Chern character form of a holomorphic Hermitian vector bundle (E,h) over a complex manifold X in a local holomorphic frame. First, we use the descent equations arising in the double complex of (p,q)-forms on X and find explicit degree decomposition of the Cher…
In math.SG/0605587, we studied Yang-Mills functional on the space of connections on a principal G_R-bundle over a closed, connected, nonorientable surface, where G_R is any compact connected Lie group. In this sequel, we generalize the discussion in "The Yang-Mills equations over Riemann surfaces" by Atiyah and Bott, a…
We determine the explicit transformation under duality of generic configurations of four flags in $\PGL(3,\bC)$ in cross-ratio coordinates. As an application we prove invariance under duality of an invariant in the Bloch group obtained from decorated triangulations of 3-manifolds.
The paper studies Kähler manifolds with partially semi-positive curvature and rational connectedness.
problem Analyzing compact Kähler manifolds with partially semi-positive curvature and rational connectedness.
method Proving rational connectedness for manifolds with BC-p positive tangent bundles, and applying these results to curvature conditions. result Confirming a conjecture and generalizing results on rational connectedness and curvature conditions.