Research output
Publications
- 2026
- Recently accepted publication
Mixture-trained merging for unified multi-objective models
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
Competing event models: next event prediction under interventions
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
Context-aware generative imputation for robust multimodal learning in missing modality scenarios
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
Coarse-to-fine compositional diffusion for long-horizon planning
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
Generative active learning via Bayesian acquisition for improving the efficiency of synthetic data
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
SAVE: sparsity-aware influence estimation for vocabulary-expanded LLMs
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
SKIM: pruning large language model agents via selective knowledge informed masking
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear - Recently accepted publication
Example-based spatial guidance for training-free concept erasure in diffusion models
Neural Information Processing Systems (NeurIPS), 2026
ConferenceNeurIPSTo appear A model-free universal AI
Conference on Uncertainty in Artificial Intelligence (UAI), 2026
ConferenceUAITo appearOCNR: stabilizing self-play by mitigating iteration-collapse with one-class novelty rewards
International Conference on Machine Learning (ICML), 2026
ConferenceICMLConfidence is not universal: task-dependent calibration and emergent behavior in LLMs
International Conference on Machine Learning (ICML), 2026
ConferenceICMLFrom drift to coherence: stabilizing beliefs in LLMs
International Conference on Machine Learning (ICML), 2026
ConferenceICMLFunctional adjoint sampler: scalable sampling on infinite dimensional spaces
International Conference on Machine Learning (ICML), 2026
ConferenceICMLRiemannian diffusion models on general manifolds via physics-informed neural networks
International Conference on Machine Learning (ICML), 2026
ConferenceICMLASCG: adaptive spatial classifier guidance for surgical concept suppression in diffusion models
CVPR 2026 Workshop — Synthetic & Adversarial Forensics (SAFE)
Symposium & WorkshopStochastic optimal control for continuous-time fMRI representation learning
International Conference on Learning Representations (ICLR), 2026
ConferenceICLRSoft equivariance regularization for invariant self-supervised learning
International Conference on Learning Representations (ICLR), 2026
ConferenceICLRForestPersons: a large-scale dataset for under-canopy missing person detection
International Conference on Learning Representations (ICLR), 2026
ConferenceICLRBridging the missing-modality gap: improving text-only calibration of vision language models
ICLR 2026 Workshop — Trustworthy AI
Symposium & WorkshopMitigating legibility tax with decoupled prover-verifier games
ICLR 2026 Workshop — Trustworthy AI
Symposium & WorkshopPermutation-symmetrized diffusion for unconditional molecular generation
ICLR 2026 Workshop — DeLTa
Symposium & WorkshopPreventing representation collapse in latent prediction via context-conditional alignment under missing modalities
ICLR 2026 Workshop — Re-Align
Symposium & Workshop- 2025
Axial neural networks for dimension-free foundation models
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSSpotlightCost-sensitive freeze-thaw Bayesian optimization for efficient hyperparameter tuning
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSFedSVD: adaptive orthogonalization for private federated learning with LoRA
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSCompact memory for continual logistic regression
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSTest time scaling for neural processes
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSPANGEA: projection-based augmentation with non-relevant general data for enhanced domain adaptation in LLMs
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSReliable decision‑making via calibration‑oriented retrieval‑augmented generation
Neural Information Processing Systems (NeurIPS), 2025
ConferenceNeurIPSInfinite dimensional adjoint sampler: scalable sampling on function spaces
NeurIPS 2025 Workshop — Frontiers in Probabilistic Inference
Symposium & WorkshopPosterImproving constrained language generation via self-distilled twisted sequential Monte Carlo
NeurIPS 2025 Workshop — Frontiers in Probabilistic Inference
Symposium & WorkshopStarFT: robust fine-tuning of zero-shot models via spuriosity alignment
International Joint Conference on Artificial Intelligence (IJCAI), 2025
ConferenceIJCAIBayesian neural scaling laws extrapolation with prior-fitted networks
International Conference on Machine Learning (ICML), 2025
ConferenceICMLEnsemble distribution distillation via flow matching
International Conference on Machine Learning (ICML), 2025
ConferenceICMLActive learning with selective time-step acquisition for PDEs
International Conference on Machine Learning (ICML), 2025
ConferenceICMLVerbalized confidence triggers self-verification: emergent behavior without explicit reasoning supervision
ICML 2025 Workshop — R2-FM
Symposium & WorkshopParameter expanded stochastic gradient Markov chain Monte Carlo
International Conference on Learning Representations (ICLR), 2025
ConferenceICLRDimension agnostic neural processes
International Conference on Learning Representations (ICLR), 2025
ConferenceICLRVariational Bayesian pseudo-coreset
International Conference on Learning Representations (ICLR), 2025
ConferenceICLRAmortized control of continuous state space Feynman-Kac model for irregular time series
International Conference on Learning Representations (ICLR), 2025
ConferenceICLROralLearning diverse attacks on large language models for robust red-teaming and safety tuning
International Conference on Learning Representations (ICLR), 2025
ConferenceICLRHarmAug: effective data augmentation for knowledge distillation of safety guard models
International Conference on Learning Representations (ICLR), 2025
ConferenceICLROver-parameterised shallow neural networks with asymmetrical node scaling: global convergence guarantees and feature learning
Transactions on Machine Learning Research (TMLR), February 2025
JournalTMLR- 2024
Model fusion through Bayesian optimization in language model fine-tuning
Neural Information Processing Systems (NeurIPS), 2024
ConferenceNeurIPSSpotlightEx uno pluria: insights on ensembling in low precision number systems
Neural Information Processing Systems (NeurIPS), 2024
ConferenceNeurIPSLearning infinitesimal generators of continuous symmetries from data
Neural Information Processing Systems (NeurIPS), 2024
ConferenceNeurIPSStochastic optimal control for diffusion bridges in function spaces
Neural Information Processing Systems (NeurIPS), 2024
ConferenceNeurIPSSafeguard text-to-image diffusion models with human feedback inversion
European Conference on Computer Vision (ECCV), 2024
ConferenceECCVEfficient modeling of irregular time-series with stochastic optimal control
NeurIPS 2024 Workshop — Bayesian Decision-Making and Uncertainty (BDU)
Symposium & WorkshopPosterLearning diverse attacks on large language models for robust red-teaming and safety tuning
NeurIPS 2024 Workshop — Red Teaming GenAI
Symposium & WorkshopVariational partial group convolutions for input-aware partial equivariance of rotations and color-shifts
International Conference on Machine Learning (ICML), 2024
ConferenceICMLA simple early exiting framework for accelerated sampling in diffusion models
International Conference on Machine Learning (ICML), 2024
ConferenceICMLLearning to explore for stochastic gradient MCMC
International Conference on Machine Learning (ICML), 2024
ConferenceICMLStabilizing the training of consistency models with score guidance
ICML 2024 Workshop — Structured Probabilistic Inference & Generative Modeling
Symposium & WorkshopCost-sensitive multi-fidelity Bayesian optimization with transfer of learning curve extrapolation
AutoML 2024 Workshop
Symposium & WorkshopFast ensembling with diffusion Schrödinger bridge
International Conference on Learning Representations (ICLR), 2024
ConferenceICLRSparse weight averaging with multiple particles for iterative magnitude pruning
International Conference on Learning Representations (ICLR), 2024
ConferenceICLRLipsum-FT: robust fine-tuning of zero-shot models using random text guidance
International Conference on Learning Representations (ICLR), 2024
ConferenceICLREnhancing transfer learning with flexible nonparametric posterior sampling
International Conference on Learning Representations (ICLR), 2024
ConferenceICLRSelf-supervised dataset distillation for transfer learning
International Conference on Learning Representations (ICLR), 2024
ConferenceICLRLearning dynamic brain connectome with graph transformers for psychiatric diagnosis classification
IEEE International Symposium on Biomedical Imaging (ISBI) 2024
Symposium & WorkshopOralSpear and shield: adversarial attacks and defense methods for model-based link prediction on continuous-time dynamic graphs
Association for the Advancement of Artificial Intelligence (AAAI), 2024
ConferenceAAAI- 2023
Function space Bayesian pseudocoreset for Bayesian neural networks
Neural Information Processing Systems (NeurIPS), 2023
ConferenceNeurIPSA generative self-supervised framework using functional connectivity in fMRI data
NeurIPS 2023 Workshop — Temporal Graph Learning
Symposium & WorkshopLarge-scale graph representation learning of dynamic brain connectome with transformers
NeurIPS 2023 Workshop — Temporal Graph Learning
Symposium & WorkshopDeep neural networks with dependent weights: Gaussian process mixture limit, heavy tails, sparsity and compressibility
Journal of Machine Learning Research, September 2023
JournalJMLRA unified construction for series representations and finite approximations of completely random measures
Bernoulli, August 2023
JournalBernoulliProbabilistic imputation for time-series classification with missing data
International Conference on Machine Learning (ICML), 2023
ConferenceICMLTraversing between modes in function space for fast ensembling
International Conference on Machine Learning (ICML), 2023
ConferenceICMLRegularizing towards soft equivariance under mixed symmetries
International Conference on Machine Learning (ICML), 2023
ConferenceICMLScalable set encoding with universal mini-batch consistency and unbiased full set gradient approximation
International Conference on Machine Learning (ICML), 2023
ConferenceICMLEarly exiting for accelerated inference in diffusion models
ICML 2023 Workshop — Structured Probabilistic Inference & Generative Modeling
Symposium & WorkshopFunction space Bayesian pseudocoreset for Bayesian neural networks
ICML 2023 Workshop — Structured Probabilistic Inference & Generative Modeling
Symposium & WorkshopTowards safe self-distillation of internet-scale text-to-image diffusion models
ICML 2023 Workshop — Challenges in Deployable Generative AI
Symposium & WorkshopMartingale posterior neural processes
International Conference on Learning Representations (ICLR), 2023
ConferenceICLRSpotlightDecoupled training for long-tailed classification with stochastic representations
International Conference on Learning Representations (ICLR), 2023
ConferenceICLRA simple yet powerful deep active learning with snapshot ensembles
International Conference on Learning Representations (ICLR), 2023
ConferenceICLRSelf-distillation for further pre-training of transformers
International Conference on Learning Representations (ICLR), 2023
ConferenceICLRExploring the role of mean teachers in self-supervised masked auto-encoders
International Conference on Learning Representations (ICLR), 2023
ConferenceICLRModeling uplift from observational time-series in continual scenarios
AAAI 2023 Bridge — Continual Causality
Symposium & WorkshopOral- 2022
On divergence measures for Bayesian pseudocoresets
Neural Information Processing Systems (NeurIPS), 2022
ConferenceNeurIPSSet-based meta-interpolation for few-task meta-learning
Neural Information Processing Systems (NeurIPS), 2022
ConferenceNeurIPSThe Normal-Generalised Gamma-Pareto process: A novel pure-jump Lévy process with flexible tail and jump-activity properties
Bayesian Analysis, December 2022
JournalBayesian AnalysisFine-tuning diffusion models with limited data
NeurIPS 2022 Workshop — Score-Based Methods
Symposium & WorkshopBenefits of stochastic weight averaging in developing neural network radiation scheme for numerical weather prediction
Journal of Advances in Modeling Earth Systems, October 2022
JournalJAMESImproving ensemble distillation with weight averaging and diversifying perturbation
International Conference on Machine Learning (ICML), 2022
ConferenceICMLSet based stochastic subsampling
International Conference on Machine Learning (ICML), 2022
ConferenceICMLScale mixtures of neural network Gaussian processes
International Conference on Learning Representations (ICLR), 2022
ConferenceICLRSequential Reptile: inter-task gradient alignment for multilingual learning
International Conference on Learning Representations (ICLR), 2022
ConferenceICLRMeta learning low rank covariance factors for energy-based deterministic uncertainty
International Conference on Learning Representations (ICLR), 2022
ConferenceICLR- 2021
Diversity matters when learning from ensembles
Neural Information Processing Systems (NeurIPS), 2021
ConferenceNeurIPSMini-batch consistent slot set encoder for scalable set encoding
Neural Information Processing Systems (NeurIPS), 2021
ConferenceNeurIPSAdaptive strategy for resetting a non-stationary Markov chain during learning via joint stochastic optimization
Symposium on Advances in Approximate Bayesian Inference (AABI) 2021
Symposium & WorkshopA multi-mode modulator for multi-domain few-shot classification
International Conference on Computer Vision (ICCV), 2021
ConferenceICCVLearning to perturb word embeddings for out-of-distribution QA
Association for Computational Linguistics (ACL), 2021
ConferenceACLAdversarial purification with score-based generative models
International Conference on Machine Learning (ICML), 2021
ConferenceICMLSetVAE: learning hierarchical composition for generative modeling of set-structured data
Conference on Computer Vision and Pattern Recognition (CVPR), 2021
ConferenceCVPR- 2020
Bootstrapping neural processes
Neural Information Processing Systems (NeurIPS), 2020
ConferenceNeurIPSCost-effective interactive attention learning with neural attention processes
International Conference on Machine Learning (ICML), 2020
ConferenceICMLDeep mixed effect model using Gaussian processes: a personalized and reliable prediction for healthcare
Association for the Advancement of Artificial Intelligence (AAAI), 2020
ConferenceAAAI- 2019
Towards deep amortized clustering
NeurIPS 2019 Workshop — Sets & Partitions
Symposium & WorkshopContributed TalkGraph embedding VAE: a permutation invariant model of graph structure
NeurIPS 2019 Workshop — Graph Representation Learning
Symposium & WorkshopBeyond the Chinese restaurant and Pitman-Yor processes: statistical models with double power-law behavior
International Conference on Machine Learning (ICML), 2019
ConferenceICMLLong OralSet transformer: a framework for attention-based permutation-invariant neural networks
International Conference on Machine Learning (ICML), 2019
ConferenceICMLLearning to propagate labels: transductive propagation network for few-shot learning
International Conference on Learning Representations (ICLR), 2019
ConferenceICLRA Bayesian model for sparse graphs with flexible degree distribution and overlapping community structure
International Conference on Artificial Intelligence and Statistics (AISTATS), 2019
ConferenceAISTATSOral- 2018
Uncertainty-aware attention for reliable interpretation and prediction
Neural Information Processing Systems (NeurIPS), 2018
ConferenceNeurIPSDropmax: adaptive variational softmax
Neural Information Processing Systems (NeurIPS), 2018
ConferenceNeurIPS- 2017
Bayesian inference on random simple graphs with power law degree distributions
International Conference on Machine Learning (ICML), 2017
ConferenceICML- 2016
Finite-dimensional BFRY priors and variational Bayesian inference for power law models
Neural Information Processing Systems (NIPS; NeurIPS), 2016
ConferenceNeurIPS- 2015
Tree-guided MCMC inference for normalized random measure mixture models
Neural Information Processing Systems (NIPS; NeurIPS), 2015
ConferenceNeurIPSBayesian hierarchical clustering with exponential family: small-variance asymptotics and reducibility
International Conference on Artificial Intelligence and Statistics (AISTATS), 2015
ConferenceAISTATS- 2014
Incremental tree-based inference with dependent normalized random measures
International Conference on Artificial Intelligence and Statistics (AISTATS), 2014
ConferenceAISTATS- 2012
Online video segmentation by Bayesian split-merge clustering
European Conference on Computer Vision (ECCV), 2012
ConferenceECCV