Search Results for author:
Found 50 papers, 23 papers with code
Sculpting Subspaces: Constrained Full Fine-Tuning in LLMs for Continual Learning
Continual learning in large language models (LLMs) is prone to catastrophic forgetting, where adapting to new tasks significantly degrades performance on previously learned ones.
SQuat: Subspace-orthogonal KV Cache Quantization
The key-value (KV) cache accelerates LLMs decoding by storing KV tensors from previously generated tokens.
Activation-Informed Merging of Large Language Models
Model merging, a method that combines the parameters and embeddings of multiple fine-tuned large language models (LLMs), offers a promising approach to enhance model performance across various tasks while maintaining computational efficiency.
A Probabilistic Inference Approach to Inference-Time Scaling of LLMs using Particle-Based Monte Carlo Methods
In this paper, we instead cast inference-time scaling as a probabilistic inference task and leverage sampling-based techniques to explore the typical set of the state distribution of a state-space model with an approximate likelihood, rather than optimize for its mode directly.
Unveiling the Secret Recipe: A Guide For Supervised Fine-Tuning Small LLMs
The rise of large language models (LLMs) has created a significant disparity: industrial research labs with their computational resources, expert teams, and advanced infrastructures, can effectively fine-tune LLMs, while individual developers and small organizations face barriers due to limited resources.
Dr. SoW: Density Ratio of Strong-over-weak LLMs for Reducing the Cost of Human Annotation in Preference Tuning
In this paper, we introduce Dr. SoW (Density Ratio of Strong over Weak) a cost-effective method that eliminates the reliance for human annotation by leveraging off-the-shelf LLMs for preference data annotation.
DICE: Discrete Inversion Enabling Controllable Editing for Multinomial Diffusion and Masked Generative Models
Discrete diffusion models have achieved success in tasks like image generation and masked language modeling but face limitations in controlled content editing.
Urban context and delivery performance: Modelling service time for cargo bikes and vans across diverse urban environments
We introduce two datasets that allow for in-depth analysis and modelling of service times of cargo bikes and use existing datasets to reason about differences in delivery performance across vehicle types.
Spectrum-Aware Parameter Efficient Fine-Tuning for Diffusion Models
We introduce Spectral Orthogonal Decomposition Adaptation (SODA), which balances computational efficiency and representation capacity.
LInK: Learning Joint Representations of Design and Performance Spaces through Contrastive Learning for Mechanism Synthesis
Moreover, we introduce a significantly more challenging benchmark, named LINK ABC, which involves synthesizing linkages that trace the trajectories of English capital alphabets, an inverse design benchmark task that existing methods struggle with due to large nonlinearities and tiny feasible space.
Differentially Private Synthetic Data Generation for Relational Databases
Existing differentially private (DP) synthetic data generation mechanisms typically assume a single-source table.
Value Augmented Sampling for Language Model Alignment and Personalization
Aligning Large Language Models (LLMs) to cater to different human preferences, learning new skills, and unlearning harmful behavior is an important problem.
LAB: Large-Scale Alignment for ChatBots
This work introduces LAB (Large-scale Alignment for chatBots), a novel methodology designed to overcome the scalability challenges in the instruction-tuning phase of large language model (LLM) training.
Curiosity-driven Red-teaming for Large Language Models
To probe when an LLM generates unwanted content, the current paradigm is to recruit a \textit{red team} of human testers to design input prompts (i. e., test cases) that elicit undesirable responses from LLMs.
Learning to Deliver: a Foundation Model for the Montreal Capacitated Vehicle Routing Problem
In this paper, we present the Foundation Model for the Montreal Capacitated Vehicle Routing Problem (FM-MCVRP), a novel Deep Learning (DL) model that approximates high-quality solutions to a variant of the Capacitated Vehicle Routing Problem (CVRP) that characterizes many real-world applications.
Private Synthetic Data Meets Ensemble Learning
When machine learning models are trained on synthetic data and then deployed on real data, there is often a performance drop due to the distribution shift between synthetic and real data.
Beyond Uniform Sampling: Offline Reinforcement Learning with Imbalanced Datasets
We argue this is due to an assumption made by current offline RL algorithms of staying close to the trajectories in the dataset.
Identifiability Guarantees for Causal Disentanglement from Soft Interventions
Causal disentanglement aims to uncover a representation of data using latent variables that are interrelated through a causal model.
Constraining Generative Models for Engineering Design with Negative Data
Our negative-data generative model (NDGM) formulation easily outperforms classic models, generating 1/6 as many constraint-violating samples using 1/8 as much data in certain problems.
A Probabilistic Framework for Modular Continual Learning
Comparing to a wide range of approaches, we show that PICLE is the first modular CL algorithm to achieve perceptual, few-shot and latent transfer while scaling well to large search spaces, outperforming previous state-of-the-art modular CL approaches on long problem sequences.
Improving Tuning-Free Real Image Editing with Proximal Guidance
Null-text inversion (NTI) optimizes null embeddings to align the reconstruction and inversion trajectories with larger CFG scales, enabling real image editing with cross-attention control.
Estimating the Density Ratio between Distributions with High Discrepancy using Multinomial Logistic Regression
We show that if these auxiliary densities are constructed such that they overlap with $p$ and $q$, then a multi-class logistic regression allows for estimating $\log p/q$ on the domain of any of the $K+2$ distributions and resolves the distribution shift problems of the current state-of-the-art methods.
Constructive Assimilation: Boosting Contrastive Learning Performance through View Generation Strategies
Recently, several attempts have been made to replace such domain-specific, human-designed transformations with generated views that are learned.
Multi-Symmetry Ensembles: Improving Diversity and Generalization via Opposing Symmetries
In this work, we present Multi-Symmetry Ensembles (MSE), a framework for constructing diverse ensembles by capturing the multiplicity of hypotheses along symmetry axes, which explore the hypothesis space beyond stochastic perturbations of model weights and hyperparameters.
Beyond Statistical Similarity: Rethinking Metrics for Deep Generative Models in Engineering Design
This paper doubles as a review and a practical guide to evaluation metrics for deep generative models (DGMs) in engineering design.
Modelling the performance of delivery vehicles across urban micro-regions to accelerate the transition to cargo-bike logistics
Light goods vehicles (LGV) used extensively in the last mile of delivery are one of the leading polluters in cities.
On the Importance of Calibration in Semi-supervised Learning
State-of-the-art (SOTA) semi-supervised learning (SSL) methods have been highly successful in leveraging a mix of labeled and unlabeled data by combining techniques of consistency regularization and pseudo-labeling.
LINKS: A dataset of a hundred million planar linkage mechanisms for data-driven kinematic design
LINKS is made up of various components including 100 million mechanisms, the simulation data for each mechanism, normalized paths generated by each mechanism, a curated set of paths, the code used to generate the data and simulate mechanisms, and a live web demo for interactive design of linkage mechanisms.
A Bayesian-Symbolic Approach to Reasoning and Learning in Intuitive Physics
In this paper, we propose a Bayesian-symbolic framework (BSP) for physical reasoning and learning that is close to human-level sample-efficiency and accuracy.
Targeted Neural Dynamical Modeling
These approaches, however, are limited in their ability to capture the underlying neural dynamics (e. g. linear) and in their ability to relate the learned dynamics back to the observed behaviour (e. g. no time lag).
Equivariant Contrastive Learning
In state-of-the-art self-supervised learning (SSL) pre-training produces semantically good representations by encouraging them to be invariant under meaningful transformations prescribed from human knowledge.
Equivariant Self-Supervised Learning: Encouraging Equivariance in Representations
In state-of-the-art self-supervised learning (SSL) pre-training produces semantically good representations by encouraging them to be invariant under meaningful transformations prescribed from human knowledge.
Scaling Densities For Improved Density Ratio Estimation
As such, estimating density ratios accurately using only samples from $p$ and $q$ is of high significance and has led to a flurry of recent work in this direction.
A Bayesian-Symbolic Approach to Learning and Reasoning for Intuitive Physics
As such, learning the laws is then reduced to symbolic regression and Bayesian inference methods are used to obtain the distribution of unobserved properties.
not-so-big-GAN: Generating High-Fidelity Images on Small Compute with Wavelet-based Super-Resolution
First, we generate images in low-frequency bands by training a sampler in the wavelet domain.
Improving the Reconstruction of Disentangled Representation Learners via Multi-Stage Modeling
Current autoencoder-based disentangled representation learning methods achieve disentanglement by penalizing the (aggregate) posterior to encourage statistical independence of the latent factors.
not-so-BigGAN: Generating High-Fidelity Images on Small Compute with Wavelet-based Super-Resolution
First, we generate images in low-frequency bands by training a sampler in the wavelet domain.
Sequential Transfer Machine Learning in Networks: Measuring the Impact of Data and Neural Net Similarity on Transferability
As the number of data sets in business networks grows and not every neural net transfer is successful, indicators are needed for its impact on the target performance-its transferability.
CZ-GEM: A FRAMEWORK FOR DISENTANGLED REPRESENTATION LEARNING
In this work, we tackle a slightly more intricate scenario where the observations are generated from a conditional distribution of some known control variate and some latent noise variate.
SimVAE: Simulator-Assisted Training forInterpretable Generative Models
Then an inference network (encoder)is trained to invert the decoder.
BreGMN: scaled-Bregman Generative Modeling Networks
Well-definedness of f-divergences, however, requires the distributions of the data and model to overlap completely in every time step of training.
Scalable Spike Source Localization in Extracellular Recordings using Amortized Variational Inference
Determining the positions of neurons in an extracellular recording is useful for investigating functional properties of the underlying neural circuitry.
Fast and Scalable Bayesian Deep Learning by Weight-Perturbation in Adam
Uncertainty computation in deep learning is essential to design robust and reliable systems.
Generative Ratio Matching Networks
In this work, we take their insight of using kernels as fixed adversaries further and present a novel method for training deep generative models that does not involve saddlepoint optimization.
Variational Inference In Pachinko Allocation Machines
The Pachinko Allocation Machine (PAM) is a deep topic model that allows representing rich correlation structures among topics by a directed acyclic graph over topics.
HOUDINI: Lifelong Learning as Program Synthesis
We present a neurosymbolic framework for the lifelong learning of algorithmic tasks that mix perception and procedural reasoning.
VEEGAN: Reducing Mode Collapse in GANs using Implicit Variational Learning
Deep generative models provide powerful tools for distributions over complicated manifolds, such as those of natural images.
Autoencoding Variational Inference For Topic Models
A promising approach to address this problem is autoencoding variational Bayes (AEVB), but it has proven diffi- cult to apply to topic models in practice.
Ranked #6 on
Topic Models
on 20NewsGroups
Clustering with a Reject Option: Interactive Clustering as Bayesian Prior Elicitation
A good clustering can help a data analyst to explore and understand a data set, but what constitutes a good clustering may depend on domain-specific and application-specific criteria.
Clustering with a Reject Option: Interactive Clustering as Bayesian Prior Elicitation
A good clustering can help a data analyst to explore and understand a data set, but what constitutes a good clustering may depend on domain-specific and application-specific criteria.
