Search for: All records

We address the problem of active logistic regression in the realizable setting. It is well known that active learning can require exponentially fewer label queries compared to passive learning, in some cases using $$\log \frac{1}{\eps}$$ rather than $$\poly(1/\eps)$$ labels to get error $$\eps$$ larger than the optimum. We present the first algorithm that is polynomially competitive with the optimal algorithm on every input instance, up to factors polylogarithmic in the error and domain size. In particular, if any algorithm achieves label complexity polylogarithmic in $$\eps$$, so does ours. Our algorithm is based on efficient sampling and can be extended to learn more general class of functions. We further support our theoretical results with experiments demonstrating performance gains for logistic regression compared to existing active learning algorithms. 

Semantic communication is of crucial importance for the next-generation wireless communication networks. The existing works have developed semantic communication frameworks based on deep learning. However, systems powered by deep learning are vulnerable to threats such as backdoor attacks and adversarial attacks. This paper delves into backdoor attacks targeting deep learning-enabled semantic communication systems. Since current works on backdoor attacks are not tailored for semantic communication scenarios, a new backdoor attack paradigm on semantic symbols (BASS) is introduced, based on which the corresponding defense measures are designed. Specifically, a training framework is proposed to prevent BASS. Additionally, reverse engineering-based and pruning-based defense strategies are designed to protect against backdoor attacks in semantic communication. Simulation results demonstrate the effectiveness of both the proposed attack paradigm and the defense strategies. 

Land surface temperature (LST) derived from satellite observations and weather modeling has been widely used for investigating Earth surface-atmosphere energy exchange and radiation budget. However, satellite-derived LST has a trade-off between spatial and temporal resolutions and missing observations caused by clouds, while there are limitations such as potential bias and expensive computation in model calibration and simulation for weather modeling. To mitigate those limitations, we proposed a WRFM framework to estimate LST at a spatial resolution of 1 km and temporal resolution of an hour by integrating the Weather Research and Forecasting (WRF) model and MODIS satellite data using the morphing technique. We tested the framework in eight counties, Iowa, USA, including urban and rural areas, to generate hourly LSTs from June 1st to August 31st, 2019, at a 1 km resolution. Upon evaluation with in-situ LST measurements, our WRFM framework has demonstrated its ability to capture hourly LSTs under both clear and cloudy conditions, with a root mean square error (RMSE) of 2.63 K and 3.75 K, respectively. Additionally, the assessment with satellite LST observations has shown that the WRFM framework can effectively reduce the bias magnitude in LST from the WRF simulation, resulting in a reduction of the average RMSE over the study area from 4.34 K (daytime) and 4.12 K (nighttime) to 2.89 K (daytime) and 2.75 K (nighttime), respectively, while still capturing the hourly patterns of LST. Overall, the WRFM is effective in integrating the complementary advantages of satellite observations and weather modeling and can generate LSTs with high spatiotemporal resolutions in areas with complex landscapes (e.g., urban). 

Large Vision-Language Models (LVLMs) have made substantial progress by integrating pre-trained large language models (LLMs) and vision models through instruction tuning. Despite these advancements, LVLMs often exhibit the hallucination phenomenon, where generated text responses appear linguistically plausible but contradict the input image, indicating a misalignment between image and text pairs. This misalignment arises because the model tends to prioritize textual information over visual input, even when both the language model and visual representations are of high quality. Existing methods leverage additional models or human annotations to curate preference data and enhance modality alignment through preference optimization. These approaches are resource-intensive and may not effectively reflect the target LVLM's preferences, making the curated preferences easily distinguishable. Our work addresses these challenges by proposing the Calibrated Self-Rewarding (CSR) approach, which enables the model to self-improve by iteratively generating candidate responses, evaluating the reward for each response, and curating preference data for fine-tuning. In the reward modeling, we employ a step-wise strategy and incorporate visual constraints into the self-rewarding process to place greater emphasis on visual input. Empirical results demonstrate that CSR significantly enhances performance and reduces hallucinations across twelve benchmarks and tasks, achieving substantial improvements over existing methods by 7.62%. Our empirical results are further supported by rigorous theoretical analysis, under mild assumptions, verifying the effectiveness of introducing visual constraints into the self-rewarding paradigm. Additionally, CSR shows compatibility with different vision-language models and the ability to incrementally improve performance through iterative fine-tuning.