An official website of the United States government
Object detection plays a pivotal in autonomous driving by enabling the vehicles to perceive and comprehend their environment, thereby making informed decisions for safe navigation. Camera data provides rich visual context and object recognition, while LiDAR data offers precise distance measurements and 3D mapping. Multi-modal object detection models are gaining prominence in incorporating these data types, which is essential for the comprehensive perception and situational awareness needed in autonomous vehicles. Although graphics processing units (GPUs) and field-programmable gate arrays (FPGAs) are promising hardware options for this application, the complex knowledge required to efficiently adapt and optimize multi-modal detection models for FPGAs presents a significant barrier to their utilization on this versatile and efficient platform. In this work, we evaluate the performance of camera and LiDARbased detection models on GPU and FPGA hardware, aiming to provide a specialized understanding for translating multi-modal detection models to suit the unique architecture of heterogeneous hardware platforms in autonomous driving systems. We focus on critical metrics from both system and model performance aspects. Based on our quantitative implications, we propose foundational insights and guidance for the design of camera and LiDAR-based multi-modal detection models on diverse hardware platforms.
more »
« less
This content will become publicly available on July 14, 2026
BalancEdit: Dynamically Balancing the Generality-Locality Trade-off in Multi-modal Model Editing
Large multi-modal models inevitably decay over time as facts update and previously learned information becomes outdated. Traditional approaches such as fine-tuning are often impractical for updating these models due to their size and complexity. Instead, direct knowledge editing within the models presents a more viable solution. Current model editing techniques, however, typically overlook the unique influence ranges of different facts, leading to compromised model performance in terms of both generality and locality. To address this issue, we introduce the concept of the generality-locality trade-off in multi-modal model editing. We develop a new model editing dataset named OKEDIT, specifically designed to effectively evaluate this trade-off. Building on this foundation, we propose BalancEdit, a novel method for balanced model editing that dynamically achieves an optimal balance between generality and locality. BalancEdit utilizes a unique mechanism that generates both positive and negative samples for each fact to accurately determine its influence scope and incorporates these insights into the model's latent space using a discrete, localized codebook of edits, without modifying the underlying model weights. To our knowledge, this is the first approach explicitly addressing the generality-locality trade-off in multi-modal model editing. Our comprehensive results confirm the effectiveness of BalancEdit, demonstrating minimal trade-offs while maintaining robust editing capabilities.
more »
« less
- PAR ID:
- 10615786
- Publisher / Repository:
- The Forty-Second International Conference on Machine Learning
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Modern language models have the capacity to store and use immense amounts of knowledge about real-world entities, but it remains unclear how to update such knowledge stored in model parameters. While prior methods for updating knowledge in LMs successfully inject atomic facts, updated LMs fail to make inferences based on injected facts. In this work, we demonstrate that a context distillation-based approach can both impart knowledge about entities and propagate that knowledge to enable broader inferences. Our approach consists of two stages: transfer set generation and distillation on the transfer set. We first generate a transfer set by prompting a language model to generate continuations from the entity definition. Then, we update the model parameters so that the distribution of the LM (the student) matches the distribution of the LM conditioned on the definition (the teacher) on the transfer set. Our experiments demonstrate that this approach is more effective at propagating knowledge updates than fine-tuning and other gradient-based knowledge-editing methods. Moreover, it does not compromise performance in other contexts, even when injecting the definitions of up to 150 entities at once.more » « less
-
Memory Editing (ME) has emerged as an efficient method to modify erroneous facts or inject new facts into Large Language Models (LLMs). Two mainstream ME methods exist: parameter-modifying ME and parameter-preserving ME (integrating extra modules while preserving original parameters). Regrettably, previous studies on ME evaluation have two critical limitations: (i) evaluating LLMs with single edit only, neglecting the need for continuous editing, and (ii) evaluations focusing solely on basic factual triples, overlooking broader LLM capabilities like logical reasoning and reading understanding. This study addresses these limitations with contributions threefold: (i) We explore how ME affects a wide range of fundamental capabilities of LLMs under sequential editing. Experimental results reveal an intriguing phenomenon: Most parameter-modifying ME consistently degrade performance across all tasks after a few sequential edits. In contrast, parameter-preserving ME effectively maintains LLMs’ fundamental capabilities but struggles to accurately recall edited knowledge presented in a different format. (ii) We extend our evaluation to different editing settings, such as layers to edit, model size, instruction tuning, etc. Experimental findings indicate several strategies that can potentially mitigate the adverse effects of ME. (iii) We further explain why parameter-modifying damages LLMs from three dimensions: parameter changes after editing, language modeling capability, and the in-context learning capability. Our in-depth study advocates more careful use of ME in real-world scenarios.more » « less
-
Abstract The El Niño‐Southern Oscillation (ENSO) influences ocean wave activity across the Pacific, but its effects on island shores are modulated by local weather and selective sheltering of multi‐modal seas. Utilizing 41 years of high‐resolution wave hindcasts, we decipher the season‐ and locality‐dependent connections between ENSO and wave patterns around the Hawaiian Islands. The north and west‐facing shores, exposed to energetic northwest swells during boreal winters, experience the most pronounced ENSO‐related variability, with increased high‐surf activity during El Niño years. While the year‐round trade wind waves exhibit moderate correlation with ENSO, the basin‐wide climate influence is masked by locally accelerated trade winds in channels and around large headlands. The remarkable global‐to‐local pathway through the high‐resolution hindcast enables development of an ENSO‐based semi‐empirical wave model to statistically describe and predict severe wave conditions on vulnerable shores with potential application in coastal risk management and hazard mitigation for Pacific Islands and beyond.more » « less
-
Neural language models (LMs) represent facts about the world described by text. Sometimes these facts derive from training data (in most LMs, a representation of the word banana encodes the fact that bananas are fruits). Sometimes facts derive from input text itself (a representation of the sentence I poured out the bottle encodes the fact that the bottle became empty). We describe REMEDI, a method for learning to map statements in natural language to fact encodings in an LM’s internal representation system. REMEDI encodings can be used as knowledge editors: when added to LM hidden representations, they modify downstream generation to be consistent with new facts. REMEDI encodings may also be used as probes: when compared to LM representations, they reveal which properties LMs already attribute to mentioned entities, in some cases making it possible to predict when LMs will generate outputs that conflict with background knowledge or input text. REMEDI thus links work on probing, prompting, and LM editing, and offers steps toward general tools for fine-grained inspection and control of knowledge in LMs.more » « less
