Search Results for author:
Found 16 papers, 7 papers with code
Effectively Steer LLM To Follow Preference via Building Confident Directions
However, their capabilities are typically limited to steering the model into one of the two directions (i. e., bidirectional steering), and there has been no theoretical understanding to guarantee their performance.
Enhancing Foundation Models for Time Series Forecasting via Wavelet-based Tokenization
To address this question, we develop WaveToken, a wavelet-based tokenizer that allows models to learn complex representations directly in the space of time-localized frequencies.
Hard Constraint Guided Flow Matching for Gradient-Free Generation of PDE Solutions
Generative models that satisfy hard constraints are crucial in many scientific and engineering applications where physical laws or system requirements must be strictly respected.
Unraveling the Gradient Descent Dynamics of Transformers
While the Transformer architecture has achieved remarkable success across various domains, a thorough theoretical foundation explaining its optimization dynamics is yet to be fully developed.
Comparing and Contrasting Deep Learning Weather Prediction Backbones on Navier-Stokes and Atmospheric Dynamics
Remarkable progress in the development of Deep Learning Weather Prediction (DLWP) models positions them to become competitive with traditional numerical weather prediction (NWP) models.
Transferring Knowledge from Large Foundation Models to Small Downstream Models
Furthermore, AFT reliably translates improvement in pre-trained models into improvement in downstream performance, even if the downstream model is over $50\times$ smaller, and can effectively transfer complementary information learned by multiple pre-trained models.
Discovering Bias in Latent Space: An Unsupervised Debiasing Approach
The question-answering (QA) capabilities of foundation models are highly sensitive to prompt variations, rendering their performance susceptible to superficial, non-meaning-altering changes.
CoMM: Collaborative Multi-Agent, Multi-Reasoning-Path Prompting for Complex Problem Solving
Specifically, we prompt LLMs to play different roles in a problem-solving team, and encourage different role-play agents to collaboratively solve the target task.
Bridging Remote Sensors with Multisensor Geospatial Foundation Models
A key discovery of our research is that representations derived from natural images are not always compatible with the distinct characteristics of geospatial remote sensors, underscoring the limitations of existing representations in this field.
PipeRAG: Fast Retrieval-Augmented Generation via Algorithm-System Co-design
Retrieval-augmented generation (RAG) can enhance the generation quality of large language models (LLMs) by incorporating external token databases.
CaMML: Context-Aware Multimodal Learner for Large Models
In this work, we introduce Context-Aware MultiModal Learner (CaMML), for tuning large multimodal models (LMMs).
Ranked #141 on
Visual Question Answering
on MM-Vet
Lightweight In-Context Tuning for Multimodal Unified Models
In-context learning (ICL) involves reasoning from given contextual examples.
PreDiff: Precipitation Nowcasting with Latent Diffusion Models
We conduct empirical studies on two datasets: N-body MNIST, a synthetic dataset with chaotic behavior, and SEVIR, a real-world precipitation nowcasting dataset.
Ranked #1 on
Precipitation Forecasting
on SEVIR
Wrapped Cauchy Distributed Angular Softmax for Long-Tailed Visual Recognition
The class-wise distribution of angular representation becomes a sum of these kernels.
Towards Geospatial Foundation Models via Continual Pretraining
Geospatial technologies are becoming increasingly essential in our world for a wide range of applications, including agriculture, urban planning, and disaster response.
Defocus Deblur Microscopy via Head-to-Tail Cross-scale Fusion
Microscopy imaging is vital in biology research and diagnosis.
