United States

Vision-Language Models (VLMs), such as CLIP, have already seen widespread applications.
Researchers actively engage in further fine-tuning VLMs in safety-critical domains. 
In these domains, prediction rationality is crucial: the prediction should be correct and based on valid evidence.
Yet, for VLMs, the impact of fine-tuning on prediction rationality is seldomly investigated. 
To study this problem, we proposed two new metrics called Prediction Trustworthiness and Inference Reliability.
We conducted extensive experiments on various settings and observed some interesting phenomena.
On the one hand, we found that the well-adopted fine-tuning methods led to more correct predictions based on invalid evidence.
This potentially undermines the trustworthiness of correct predictions from fine-tuned VLMs.
On the other hand, having identified valid evidence of target objects, fine-tuned VLMs were more likely to make correct predictions.
Moreover, the findings are also consistent under distributional shifts and across various experimental settings.
We hope our research offer fresh insights to VLM fine-tuning.
We will release the source code upon acceptance.

AAAI 2025

Beyond Accuracy: On the Effects of Fine-Tuning Towards Vision-Language Model’s Prediction Rationality

interpretable

transparent

explainable

## Welcome to EMNLP 2024! 
We are excited to welcome you to one of the most prominent conferences in the field of Natural Language Processing. This year, EMNLP 2024 is being held in a hybrid format,
offering both virtual and in-person participation in beautiful Miami. Due to a record-breaking number of submissions, we've expanded the total number of accepted papers to accommodate more cutting-edge research from around the globe.
### [Conference Handbook](https://drive.google.com/file/d/1WPROgxjLAC96AJL7Ugy0tEnYm7dkrbHt/view?usp=sharing)

You are required to register for this event. **Please register [here](https://2024.emnlp.org/registration/).** The EMNLP 2024 event page on Underline will be open to public one week prior to the event.

Please register!

EMNLP 2024

EMNLP 2024 will take place in Miami, Florida from Nov 12th to Nov 16th, 2024, at the Hyatt Regency Miami Hote and on Underline for remote participants.

poster

We are pleased to announce the Thirty-Ninth AAAI Conference on Artificial Intelligence (AAAI-25), which will be held in Philadelphia, Pennsylvania at the Pennsylvania Convention Center from February 25 to March 4, 2025.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-25 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.

### [Invited Speakers](https://aaai.org/conference/aaai/aaai-25/aaai-25-invited-speakers/)

Register [here](https://aaai.org/conference/aaai/aaai-25/registration/)

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-25 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.



Large language models (LLMs) excel in language tasks, especially with supervised fine-tuning after pre-training. However, their substantial memory and computational requirements hinder practical applications. Structural pruning, which reduces less significant weight dimensions, is one solution. Yet, traditional post-hoc pruning often leads to significant performance loss, with limited recovery from further fine-tuning due to reduced capacity. Since the model fine-tuning refines the general and chaotic knowledge in pre-trained models, we aim to incorporate structural pruning with the fine-tuning, and propose the Pruning-Aware Tuning (PAT) paradigm to eliminate model redundancy while preserving the model performance to the maximum extend.
Specifically, we insert the innovative Hybrid Sparsification Modules (HSMs) between the Attention and FFN components to accordingly sparsify the upstream and downstream linear modules. The HSM comprises a lightweight operator and a globally shared trainable mask. The lightweight operator maintains a training overhead comparable to that of LoRA, while the trainable mask unifies the channels to be sparsified, ensuring structural pruning.
Additionally, we propose the Identity Loss which decouples the transformation and scaling properties of the HSMs to enhance training robustness. Extensive experiments demonstrate that PAT excels in both performance and efficiency. For example, our Llama2-7b model with a 25\% pruning ratio achieves 1.33$\times$ speedup while outperforming the LoRA-finetuned model by up to 1.26\% in accuracy with a similar training cost.

PAT: Pruning-Aware Tuning for Large Language Models

As a potential application of Vehicle-to-Everything (V2X) communication, multi-agent collaborative perception has achieved significant success in 3D object detection. While these methods have demonstrated impressive results on standard benchmarks, the robustness of such approaches in the face of complex real-world environments requires additional verification. 
To bridge this gap, we introduce the first comprehensive benchmark designed to evaluate the robustness of collaborative perception methods in the presence of natural corruptions typical of real-world environments.
Furthermore, we propose DSRC, a robustness-enhanced collaborative perception method aiming to learn Density-insensitive and Semantic-aware collaborative Representation against Corruptions. DSRC consists of two key designs: i) a semantic-guided sparse-to-dense distillation framework, which constructs multi-view dense objects painted by ground truth bounding boxes to effectively learn density-insensitive and semantic-aware collaborative representation; ii) a feature-to-point cloud reconstruction approach to better fuse critical collaborative representation across agents. To thoroughly evaluate DSRC, we conduct extensive experiments on real-world and simulated datasets. The results demonstrate that our method outperforms state-of-the-art collaborative perception methods in both clean and corrupted conditions. The code and reproducibility checklist are available in  the supplementary materials.

DSRC: Learning Density-insensitive and Semantic-aware Collaborative Representation against Corruptions

Contrastive Language-Image Pretraining (CLIP) has achieved excellent performance over a wide range of tasks. However, the effectiveness of CLIP heavily relies on a substantial corpus of pretraining data, resulting in notable consumption of computational resources. Although knowledge distillation has been widely applied in single modality models, how to efficiently expand knowledge distillation to vision-language foundation models with extensive data remains relatively unexplored. In this paper, we introduce CLIP-CID, a novel distillation mechanism that effectively transfers knowledge from a large vision-language foundation model to a smaller model. We initially propose a simple but efficient image semantic balance method to reduce transfer learning bias and improve distillation efficiency. This method filters out 43.7% of image-text pairs from the LAION400M while maintaining superior performance. After that, we leverage cluster-instance discrimination to facilitate knowledge transfer from the teacher model to the student model, thereby empowering the student model to acquire a holistic semantic comprehension of the pretraining data. Experimental results demonstrate that CLIP-CID achieves state-of-the-art performance on various downstream tasks including linear probe and zero-shot classification.

CLIP-CID: Efficient CLIP Distillation via Cluster-Instance Discrimination

Infrared (IR) spectroscopy is a fundamental technique in analytical chemistry. Recently, deep learning (DL) has drawn great interest as the modeling method of infrared spectral data. However, unlike vision or language tasks, IR spectral data modeling is faced with the problem of calibration transfer and has distinctive characteristics. Introducing the prior knowledge of IR spectroscopy could guide the DL methods to learn representations aligned with the domain-invariant characteristics of spectra, and thus improve the performance. Despite such potential, there is a notable absence of DL methods that incorporate such inductive bias. To this end, we propose Analytical-Chemistry-Informed Transformer (ACT) with two modules informed by the field knowledge in analytical chemistry. First, ACT includes learnable spectral processing inspired by chemometrics, which comprises spectral pre-processing, tokenization, and post-processing. Second, a straightforward yet effective attention mechanism, namely spectral-attention, is incorporated into ACT. Spectral-attention utilizes the intra-spectral and inter-spectral correlations to extract spectral representations. Empirical results show that ACT has achieved competitive results in 9 analytical tasks covering applications across pharmacy, chemistry, and agriculture. Compared with existing networks, ACT reduces the root mean square error of prediction (RMSEP) by more than 20\% in calibration transfer tasks. These results indicate that DL methods in IR spectroscopy could benefit from the integration of prior knowledge in analytical chemistry.

Analytical-Chemistry-Informed Transformer for Infrared Spectra Modeling

Bird's-Eye-View (BEV) representation has emerged as a mainstream paradigm for multi-view 3D object detection, demonstrating impressive perceptual capabilities. However, existing methods overlook the geometric quality of BEV representation, leaving it in a low-resolution state and failing to restore the authentic geometric information of the scene. In this paper, we identify the drawbacks of previous approaches that limit the geometric quality of BEV representation and propose Radial-Cartesian BEV Sampling (RC-Sampling), which outperforms other feature transformation methods in efficiently generating high-resolution dense BEV representations to restore fine-grained geometric information. Additionally, we design a novel In-Box Label to substitute the traditional depth label generated from the LiDAR points. This label reflects the actual geometric structure of objects rather than just their surfaces, injecting real-world geometric information into the BEV representation. In conjunction with the In-Box Label, Centroid-Aware Inner Loss (CAI Loss) is developed to capture the inner geometric structure of objects. Finally, we integrate the aforementioned modules into a novel multi-view 3D object detection framework, dubbed GeoBEV. It achieves a state-of-the-art result of 66.2% NDS on the nuScenes test set and surpasses 2.6% NDS over StreamPETR. The code and models will be made publicly available.

GeoBEV: Learning Geometric BEV Representation for Multi-view 3D Object Detection

This paper presents a novel framework for multi-armed bandit problems with side-observations and switching constraints, which arises in a range of real-world applications such as robotic. To address the challenges of effectively utilizing graph-structured observations while adhering to graph constraints, we design graph-agnostic and graph-aware algorithms tailored to this new setting. Specifically, our graph-agnostic algorithm selects nodes with the highest upper confidence bound without prior knowledge of feedback probabilities, while minimizing switching costs using offline shortest path planning and the doubling trick. If the graph structure and associated probability matrix are known, our graph-aware algorithm plans the exploration step using a linear programming approach and eliminates suboptimal nodes iteratively. We rigorously analyze the performance of our proposed algorithms, providing near-optimal minimax and instance-dependent regret upper bounds. Our analysis shows that our algorithms outperform generic reinforcement learning methods in terms of both regret and computational efficiency. Extensive numerical experiments on various types of graphs, including two real-world datasets, demonstrate the efficacy of our proposed methods and their advantages over benchmark methods in graph bandit settings.

Efficient Graph Bandit Learning with Side-Observations and Switching Constraints

Although existing Sparsely Annotated Object Detection (SAOD) approches have made progress in handling sparsely annotated environments in multispectral domain, where only some pedestrians are annotated, they still have the following limitations: (i) they lack considerations for improving the quality of pseudo-labels for missing annotations, and (ii) they rely on fixed ground truth annotations, which leads to learning only a limited range of pedestrian visual appearances in the multispectral domain. To address these issues, we propose a novel framework called Sparsely Annotated Multispectral Pedestrian Detection (SAMPD). For limitation (i), we introduce Multispectral Pedestrian-aware Adaptive Weight (MPAW) and Positive Pseudo-label Enhancement (PPE) module. Utilizing multispectral knowledge, these modules ensure the generation of high-quality pseudo-labels and enable effective learning by increasing weights for high-quality pseudo-labels based on modality characteristics. To address limitation (ii), we propose an Adaptive Pedestrian Retrieval Augmentation (APRA) module, which adaptively incorporates pedestrian patches from ground-truth and dynamically integrates high-quality pseudo-labels with the ground-truth, facilitating a more diverse learning pool of pedestrians. Extensive experimental results demonstrate that our SAMPD significantly enhances performance in sparsely annotated environments within the multispectral domain.

Multispectral Pedestrian Detection with Sparsely Annotated Label

Non-Rigid Structure-from-Motion (\textbf{NRSfM}) is a classic 3D vision problem, where a 2D sequence is taken as input to estimate the corresponding 3D sequence. Recently, the deep neural networks have greatly advanced the task of NRSfM. However, existing deep NRSfM methods still have limitations in handling the inherent sequence property and motion ambiguity associated with the NRSfM problem. In this paper, we revisit deep NRSfM from two perspectives to address the limitations of current deep NRSfM methods : \textbf{(1)} canonicalization and \textbf{(2)} sequence modeling. We propose an easy-to-implement per-sequence canonicalization method as opposed to the previous per-dataset canonicalization approaches. With this in mind, we propose a sequence modeling method that combines temporal information and subspace constraint. As a result, we have achieved a more optimal NRSfM reconstruction pipeline compared to previous efforts. The effectiveness of our method is verified by testing the sequence-to-sequence deep NRSfM pipeline with corresponding regularization modules on several commonly used datasets.

Deep Non-Rigid Structure-from-Motion Revisited: Canonicalization and Sequence Modeling

Diffusion models have exhibited substantial success in text-to-image generation. However, they often encounter challenges when dealing with complex and dense prompts involving multiple objects, attribute binding, and long descriptions. In this paper, we propose a novel framework called LLM4GEN, which enhances the semantic understanding of text-to-image diffusion models by leveraging the representation of Large Language Models (LLMs). It can be seamlessly incorporated into various diffusion models as a plug-and-play component. A specially designed Cross-Adapter Module (CAM) integrates the original text features of text-to-image models with LLM features, thereby enhancing text-to-image generation. Additionally, to facilitate and correct entity-attribute relationships in text prompts, we develop an entity-guided regularization loss to further improve generation performance. We also introduce DensePrompts, which contains 7,000 dense prompts to provide a comprehensive evaluation for the text-to-image generation task. Experiments indicate that LLM4GEN significantly improves the semantic alignment of SD1.5 and SDXL, demonstrating increases of 9.69% and 12.90% in color on T2I-CompBench, respectively. Moreover, it surpasses existing models in terms of sample quality, image-text alignment, and human evaluation.

LLM4GEN: Leveraging Semantic Representation of LLMs for Text-to-Image Generation

Deep multi-view clustering incorporating graph learning has presented tremendous potential. Most methods encounter costly square time consumption w.r.t. data size. Theoretically, anchor-based graph learning can alleviate this limitation, but related deep models mainly rely on manual discretization approaches to select anchors, which indicates that 1) the anchors are fixed during model training and 2) they may deviate from the true cluster distribution. Consequently, the unreliable anchors may corrupt clustering results. In this paper, we propose the Deep Multi-view Anchor Clustering (DMAC) model that performs clustering in linear time. Concretely, the initial anchors are intervened by the positive perturbation sampled from Gaussian distribution, such that they can be optimized with a newly designed anchor learning loss, which promotes a clear relationship between samples and anchors. Afterwards, anchor graph convolution is devised to model the cluster structure formed by the anchors, and the mutual information maximization loss is built to provide cross-view clustering guidance. In this way, the learned anchors can better represent clusters. With the optimal anchors, the full sample graph is calculated to derive a discriminative embedding for clustering. Extensive experiments on several datasets demonstrate the superior performance and efficiency of DMAC compared to state-of-the-art competitors.

Downloads

Next from AAAI 2025

PAT: Pruning-Aware Tuning for Large Language Models

Similar lecture

Vision-Language Model Fine-Tuning via Simple Parameter-Efficient Modification