United States

In various academic and professional settings, such as mathematics lectures or research presentations, it is often necessary to convey mathematical expressions orally. However, reading mathematical expressions aloud without accompanying visuals can significantly hinder comprehension, especially for those who are hearing-impaired or rely on subtitles due to language barriers. For instance, when a presenter reads Euler&#39;s Formula, current Automatic Speech Recognition (ASR) models often produce a verbose and error-prone textual description (e.g., e to the power of i x equals cosine of x plus i $\textit{side}$ of x), instead of the concise $\LaTeX{}$  format (i.e., $ e^{ix} = \cos(x) + i\sin(x) $), which hampers clear understanding and communication. To address this issue, we introduce MathSpeech, a novel pipeline that integrates ASR models with small Language Models (sLMs) to correct errors in mathematical expressions and accurately convert spoken expressions into structured $\LaTeX{}$ representations. 
Evaluated on a new dataset derived from lecture recordings, MathSpeech demonstrates$\LaTeX{}$  generation capabilities comparable to leading commercial Large Language Models (LLMs), while leveraging fine-tuned small language models of only 120M parameters.
Specifically, in terms of CER, BLEU, and ROUGE scores for $\LaTeX{}$  translation, MathSpeech demonstrated significantly superior  capabilities compared to GPT-4o. We observed a decrease in CER from 0.390 to 0.298, and higher ROUGE/BLEU scores compared to GPT-4o.

AAAI 2025

MathSpeech: Leveraging Small LMs for Accurate Conversion in Mathematical Speech-to-Formula

snlp

applications

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.



In various academic and professional settings, such as mathematics lectures or research presentations, it is often necessary to convey mathematical expressions orally. However, reading mathematical expressions aloud without accompanying visuals can significantly hinder comprehension, especially for those who are hearing-impaired or rely on subtitles due to language barriers. For instance, when a presenter reads Euler's Formula, current Automatic Speech Recognition (ASR) models often produce a verbose and error-prone textual description (e.g., e to the power of i x equals cosine of x plus i $\textit{side}$ of x), instead of the concise $\LaTeX{}$  format (i.e., $ e^{ix} = \cos(x) + i\sin(x) $), which hampers clear understanding and communication. To address this issue, we introduce MathSpeech, a novel pipeline that integrates ASR models with small Language Models (sLMs) to correct errors in mathematical expressions and accurately convert spoken expressions into structured $\LaTeX{}$ representations. 
Evaluated on a new dataset derived from lecture recordings, MathSpeech demonstrates$\LaTeX{}$  generation capabilities comparable to leading commercial Large Language Models (LLMs), while leveraging fine-tuned small language models of only 120M parameters.
Specifically, in terms of CER, BLEU, and ROUGE scores for $\LaTeX{}$  translation, MathSpeech demonstrated significantly superior  capabilities compared to GPT-4o. We observed a decrease in CER from 0.390 to 0.298, and higher ROUGE/BLEU scores compared to GPT-4o.

poster

In the domain of spatial-temporal video super-resolution, it is typically challenging to handle complex motions (includ- ing large and nonlinear motions) and varying illumination scenes due to the lack of inter-frame information. Leverag- ing the dense temporal information provided by event sig- nals appears to be a reasonable solution to this problem. Tra- ditional event-based video super-resolution methods require multiple images as input and complete the process through motion estimation and motion compensation, where the mo- tion estimation can introduce errors. When multiple images are used as input, the errors introduced by each image can lead to artifacts and blurriness. To address this issue, we pro- pose to use fewer adjacent frames and integrate dense tem- poral information from events to accomplish this task. Our method, EvSTVSR, utilizes inter-frame events to guide align- ment. We introduce a coordinate-based feature fusion upsam- pling module to achieve spatial super-resolution. Experimen- tal results demonstrate that not only does our super-resolution output outperform other time-based methods, but it also ex- hibits greater advantages in super-resolving large motions.

EvSTVSR: Event Guided Space-Time Video Super-Resolution

Shadows can originate from occlusions in both direct and indirect illumination. Although most current shadow removal research focuses on shadows caused by direct illumination, shadows from indirect illumination are often just as pervasive, particularly in indoor scenes. A significant challenge in removing shadows from indirect illumination is obtaining shadow-free images to train the shadow removal network. To overcome this challenge, we propose a novel rendering pipeline for generating shadowed and shadow-free images under direct and indirect illumination, and create a comprehensive synthetic dataset that contains over 30,000 image pairs, covering various object types and lighting conditions. We also propose an innovative shadow removal network that explicitly integrates semantic and geometric priors through concatenation and attention mechanisms. The experiments show that our method outperforms state-of-the-art shadow removal techniques and can effectively generalize to indoor and outdoor scenes under various lighting conditions, enhancing the overall effectiveness and applicability of shadow removal methods.

OmniSR: Shadow Removal Under Direct and Indirect Lighting

In long-term series forecasting (LTSF), it is imperative for models to adeptly discern and distill from historical time series data to forecast future states. Although Transformer-based models excel at capturing long-term dependencies in LTSF, their practical use is limited by issues like computational inefficiency, noise sensitivity, and overfitting on smaller datasets. Therefore, we introduce a novel time series lightweight interactive Mamba with an adaptive Fourier filter model (Affirm). Specifically, (i) we propose an adaptive Fourier filter block. This neural operator employs Fourier analysis to refine feature representation, reduces noise with learnable adaptive thresholds, and captures inter-frequency interactions using global and local semantic adaptive Fourier filters via element-wise multiplication. (ii) A dual interactive Mamba block is introduced to facilitate efficient intra-modal interactions at different granularities, capturing more detailed local features and broad global contextual information, providing a more comprehensive representation for LTSF. Extensive experiments on multiple benchmarks demonstrate that Affirm consistently outperforms existing SOTA methods, offering a superior balance of accuracy and efficiency, making it ideal for various challenging scenarios with noise levels and data sizes.

Affirm: Interactive Mamba with Adaptive Fourier Filters for Long-term Time Series Forecasting

Composing music for video is essential yet challenging, leading to a growing interest in automating music generation for video applications. Existing approaches often struggle to achieve robust music-video correspondence and generative diversity, primarily due to inadequate feature alignment methods and insufficient datasets. In this study, we present $\textbf{G}$eneral $\textbf{V}$ideo-to-$\textbf{M}$usic $\textbf{Gen}$eration model ($\textbf{GVMGen}$), designed for generating high-related music to the video input. Our model employs hierarchical attentions to extract and align video features with music in both spatial and temporal dimensions, ensuring the preservation of pertinent features while minimizing redundancy. Remarkably, our method is versatile, capable of generating multi-style music from different video inputs, even in zero-shot scenarios. We also propose an evaluation model along with two novel objective metrics for assessing video-music alignment. Additionally, we have compiled a large-scale dataset comprising diverse types of video-music pairs. Experimental results demonstrate that GVMGen surpasses previous models in terms of music-video correspondence, music quality generative diversity, and application universality.

GVMGen: A General Video-to-Music Generation Model With Hierarchical Attentions

Attribute classification is crucial for identifying specific characteristics within image regions. Vision-Language Models
(VLMs) have been particularly effective in zero-shot tasks by leveraging their general knowledge from large-scale datasets.
Recent studies demonstrate that zero-shot multi-label classification, including attribute classification, can be effectively addressed by transformer-based models with classwise queries. However, attribute classification generally involves a large number of attribute classes and that makes it difficult to maintain the model’s scalability. Additionally, poor utilization of the relationship between seen and unseen attributes leads the model to lack of generalizability. To address these issues, we propose Super-class guided transFormer (SugaFormer), a framework that leverages super-classes with Vision-Language Models (VLMs) through Super-class Query Initialization and Super-class guided Consistency Regularization for attribute classification. Superclass Query Initialization reduces the number of queries by
aligning attributes with their relevant super-classes enhancing the generalizability and scalability of the model. Super-class guided Consistency Regularization encourages features of SugaFormer to be similar to those of the VLMs by using region-specific prompts and their corresponding tokens. Our experiments and analyses demonstrate the effectiveness of SugaFormer, achieving state-of-the-art results in three widely-used attribute classification benchmarks under zero-shot, and cross-dataset transfer settings.

Super-Class Guided Transformer for Zero-Shot Attribute Classification

Weakly Supervised Semantic Segmentation (WSSS) with image-level labels typically uses Class Activation Maps (CAM) to achieve dense predictions. Recently, Vision Transformer (ViT) has provided an alternative to generate localization maps from class-patch attention. However, due to the little constraint to model such attention, we observe that the Localization Attention Maps (LAM) usually struggle with the artifact issue, i.e., patch regions with minimal semantic relevance are falsely activated by class tokens. In this work, we propose MoRe to address this issue and further explore the potential of LAM. Our findings suggest that additional regularization on class-patch attention is necessary. To this end, we first view the attention as a novel directed graph and propose the Graph Category Representation module to implicitly regularize the interaction among class-patch entities. It ensures that class tokens dynamically condense the related patch information and suppress unrelated artifacts at a graph level. Second, motivated by the observation that CAM from classification weights maintains smooth localization of objects, we devise the Localization-informed Regularization module to explicitly regularize the class-patch attention. It directly mines the token relations from CAM and further supervises the consistency between class and patch tokens in a learnable manner. Extensive experiments are conducted on PASCAL VOC and MS COCO, validating the efficiency of MoRe in addressing the artifact issue and thereby achieving state-of-the-art performance over recent single-stage and even multi-stage methods. Codes will be available.

MoRe: Class Patch Attention Needs Regularization for Weakly Supervised Semantic Segmentation

Large-scale pre-trained Vision Transformer (ViT) models have demonstrated remarkable performance on visual tasks but are computationally expensive to transfer to downstream tasks. Parameter-Efficient Fine-Tuning (PEFT) offers a promising transferring approach by updating only a subset of parameters. However, PEFT's effectiveness is hindered by discrepancies between pre-training and downstream tasks in terms of object scale and granularity. Downstream tasks often focus on finer-grained and more specialized recognition, requiring more detailed features. The diversity of feature scales of existing PEFT methods for ViT is limited. To address this, we propose a novel PEFT method named Wavelet-based multi-Scale Tuning (WST), which learns multi-scale features in a simple and efficient way. WST introduces a parallel fine-tuning patch embedding branch with a smaller patch size than the pre-trained model to capture finer-grained features. Furthermore, to handle the computational challenge from the resulting longer token sequence, WST designs wavelet fine-tuning blocks that balance both efficiency and performance. In the block, wavelet transform enables invertible and lossless down-sampling of the longer token sequence, aligning it with that of the backbone, and two lightweight linear mappings are employed to learn task-specific features. This design facilitates efficient multi-scale information exchange between the pre-trained backbone and fine-tuning branch. Extensive experiments on transfer learning demonstrate the promising performance and efficiency of our WST. Notably, our WST achieves state-of-the-art average accuracy of 78\% with only 0.08M trainable parameters on VTAB-1K vision benchmark.

WST: Wavelet-Based Multi-scale Tuning for Visual Transfer Learning

Long-tailed (LT) data distribution is common in multi-label image classification (MLC) and can significantly impact the performance of classification models. One reason is the challenge of learning unbiased instance representations (i.e. features) for imbalanced datasets. Additionally, the co-occurrence of head/tail classes within the same instance, along with complex label dependencies, introduces further challenges. In this work, we delve into this problem through the lens of neural collapse (NC). NC refers to a phenomenon where the last-layer features and classifier of a deep neural network model exhibit a simplex Equiangular Tight Frame (ETF) structure during its terminal training phase. This structure creates an optimal linearly separable state. However, this phenomenon typically occurs in balanced datasets but rarely applies to the typical imbalanced problem. To induce NC properties under Long-tailed multi-label classification (LT-MLC) conditions, we propose an approach named MLC-NC, which aims to learn high-quality data representations and improve the model’s generalization ability. Specifically, MLC-NC accounts for the fact that different labels correspond to different feature parts located in images. MLC-NC extracts class-wise features from each instance through a cross-attention mechanism. To guide the features toward the ETF structure, we introduce visual-semantic feature alignment with a fixed ETF structured label embedding, which helps to learn evenly distributed class centers. To reduce within-class feature variation, we introduce collapse calibration within a lower-dimensional feature space. To mitigate classification bias, we concatenate features and feed them into a binarized fixed ETF classifier. As an orthogonal approach to existing methods, MLC-NC can be seamlessly integrated into various frameworks. Extensive experiments on widely-used benchmarks demonstrate the effectiveness of our method.

MLC-NC: Long-Tailed Multi-Label Image Classification Through the Lens of Neural Collapse

Human Activity Recognition (HAR) aims to recognize activities by training models on massive sensor data. In real-world deployment, a crucial aspect of HAR that has been largely overlooked is that the test sets may have different distributions from training sets due to inter-subject variability including age, gender, behavioral habits, etc., which leads to poor generalization performance. One promising solution is to learn domain-invariant representations to enable a model to generalize on an unseen distribution. However, most existing methods only consider the feature-invariance of the penultimate layer for domain-invariant learning, which leads to suboptimal results. In this paper, we propose a Categorical Concept Invariant Learning (CCIL) framework for generalizable activity recognition, which introduces a concept matrix to regularize the model in the training stage by simultaneously concertrating on feature-invariance and logit-invariance. Our key idea is that the concept matrix for samples belonging to the same activity category should be similar. Extensive experiments on four public HAR benchmarks demonstrate that our CCIL substantially outperforms the state-of-the-art approaches under cross-person, cross-dataset, cross-position, and one-person-to-another settings. Code will be released.

Generalizable Sensor-Based Activity Recognition via Categorical Concept Invariant Learning

With the emergence of vision-language pre-trained models, such as CLIP, some textual prompts have been gradually introduced recently into re-identification (Re-ID) tasks to obtain considerably robust multimodal information. However, most textual descriptions based on vehicle Re-ID tasks only contain identity index words without specific words to describe vehicle view information, thereby resulting in difficulty to be widely applied in vehicle Re-ID tasks with view variations. This case inspires us to propose a CLIP-driven view-aware prompt learning framework for unsupervised vehicle Re-ID. We first design a learnable textual prompt template called view-aware context optimization (ViewCoOp) based on dynamic multi-view word embeddings, which can fully obtain the proportion and position encoding of each view in the whole vehicle body region. Subsequently, a cross-modal mutual graph is constructed to explore the connections between inter-modal and intra-modal. Each sample is treated as a graph node, which extracts textual features based on ViewCoOp and the visual features of images. Moreover, leveraging the inter-cluster and intra-cluster correlation in the bimodal clustering results in the determination of connectivity between graph node pairs. Lastly, the proposed cross-modal mutual graph method utilizes supervised information from the bimodal gap to directly fine-tune the image encoder of CLIP for downstream unsupervised vehicle Re-ID tasks. Extensive experiments verify that the proposed method is capable of effectively obtaining cross-modal description ability from multiple views.

Downloads

Next from AAAI 2025

EvSTVSR: Event Guided Space-Time Video Super-Resolution

Similar lecture

LATTE: Improving Latex Recognition for Tables and Formulae with Iterative Refinement