United States

Incremental few-shot object detection aims at detecting novel classes without forgetting knowledge of the base classes with only a few labeled training data from the novel classes. Most related prior works are on incremental object detection that rely on the availability of abundant training samples per novel class that substantially limits the scalability to real-world setting where novel data can be scarce. In this paper, we propose the Incremental-DETR that does incremental few-shot object detection via fine-tuning and self-supervised learning on the DETR object detector. To alleviate severe over-fitting with few novel class data, we first fine-tune the class-specific components of DETR with self-supervision from additional object proposals generated using Selective Search as pseudo labels. We further introduce an incremental few-shot fine-tuning strategy with knowledge distillation on the class-specific components of DETR to encourage the network in detecting novel classes without forgetting the base classes. Extensive experiments conducted on standard incremental object detection and incremental few-shot object detection settings show that our approach significantly outperforms state-of-the-art methods by a large margin.  Our source code is available at https://github.com/dongnana777/Incremental-DETR.

AAAI 2023

Incremental-DETR: Incremental Few-Shot Object Detection via Self-Supervised Learning

incremental few-shot object detection

self-supervised learning

fine-tuning

transformer

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-23 is the Thirty-Seventh AAAI Conference on Artificial Intelligence. The theme of this conference is to create collaborative bridges within and beyond AI. Like previous AAAI conferences, AAAI-23 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and two new activities: a Bridge Program and a Lab Program. Many of these activities are tailored to the theme of bridges and all are selected according to the highest standards, with additional programs for students and young researchers. 
AAAI is providing you with a conference planner, which you can use to help organize your itinerary of activities. This includes talks to attend in person, talks to attend remotely, breaks with colleagues and your site seeing activities. To access this conference planner, please go to [https://aaai-2023.takemobi.io](https://aaai-2023.takemobi.io).

In order to access this site, you need to register. If you haven't already, please register [here](https://aaai.org/Conferences/AAAI-23/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.

poster

In this paper, we address a special scenario of semi-supervised learning, where the label missing is caused by a preceding filtering mechanism, i.e., an instance can enter a subsequent process in which its label is revealed if and only if it passes the filtering mechanism. The rejected instances are prohibited to enter the subsequent labeling process due to economical or ethical reasons, making the support of the labeled and unlabeled distributions isolated from each other. In this case, semi-supervised learning approaches which rely on certain coherence of the labeled and unlabeled distribution would suffer from the consequent distribution mismatch, and hence result in poor prediction performance. In this paper, we propose a Small-Paced Self-Training framework, which iteratively discovers labeled and unlabeled instance subspaces with bounded Wasserstein distance. We theoretically prove that such a framework may achieve provably low error on the pseudo labels during learning. Experiments on both benchmark and pneumonia diagnosis tasks show that our method is effective.

Semi-Supervised Learning with Support Isolation by Small-Paced Self-Training

Compared with the image-based static facial expression recognition (SFER) task, the dynamic facial expression recognition (DFER) task based on video sequences is closer to the natural expression recognition scene. However, DFER is often more challenging. One of the main reasons is that video sequences often contain frames with different expression intensities, especially for the facial expressions in the real-world scenarios, while the images in SFER frequently present uniform and high expression intensities. Nevertheless, if the expressions with different intensities are treated equally, the features learned by the networks will have large intra-class and small inter-class differences, which are harmful to DFER. To tackle this problem, we propose the global convolution-attention block (GCA) to rescale the channels of the feature maps. In addition, we introduce the intensity-aware loss (IAL) in the training process to help the network distinguish the samples with relatively low expression intensities. Experiments on two in-the-wild dynamic facial expression datasets (i.e., DFEW and FERV39k) indicate that our method outperforms the state-of-the-art DFER approaches. The source code will be available at https://github.com/muse1998/IAL-for-Facial-Expression-Recognition.

Intensity-Aware Loss for Dynamic Facial Expression Recognition in the Wild

Contrastive learning has emerged as one of the most promising self-supervised methods. It can efficiently learn the transferable representations of samples through the instance-level discrimination task. In general, the performance of the contrastive learning method can be further improved by projecting the transferable high-dimensional representations into the low-dimensional feature space. This is because the model can learn more abstract discriminative information. However, when low-dimensional features cannot provide sufficient discriminative information to the model (e.g., the samples are very similar to each other), the existing contrastive learning method will be limited to a great extent. Therefore, in this paper, we propose a general module called the Feature Reconstruction Amplifier (FRA) for adding additional high-dimensional feature information to the model. Specifically, FRA reconstructs the low-dimensional feature embeddings with Gaussian noise vectors and projects them to a high-dimensional reconstruction space. In this reconstruction space, we can add additional feature information through the designed loss. We have verified the effectiveness of the module itself through exhaustive ablation experiments. In addition, we perform linear evaluation and transfer learning on five common visual datasets, the experimental results demonstrate that our method is superior to recent advanced contrastive learning methods.

Contrastive Learning with the Feature Reconstruction Amplifier

A {\em knockout (or single-elimination) tournament} is a format of a competition that is very popular in practice (particularly in sports, elections and decision making), and which has been extensively and intensively studied from a theoretical point of view for more than a decade. Particular attention has been devoted to the {\sc Tournament Fixing} problem, where, roughly speaking, the objective is to determine whether we can conduct the knockout tournament in a way that makes our favorite player win. Here, part of the input is a tournament graph $D$ that encodes the winner of each possible match. A sequence of papers has studied the parameterized complexity of {\sc Tournament Fixing} with respect to the {\em feedback arc set number} ($\mathbf{fas}$) of $D$. Given that this parameter yielded tractability (specifically, solvability in time  $\mathbf{fas}^{\OO(\mathbf{fas})}\cdot n^{\OO(1)}$), it has been asked explicitly and repeatedly whether {\sc Tournament Fixing} is FPT also with respect to the {\em feedback vertex set number} ($\mathbf{fvs}$) of $D$ [Ramanujan and Szeider, AAAI'17; Gupta et al., IJCAI'18; Gupta et al., IJCAI'19; Roy, PhD Thesis, 2021; Suksompong, IJCAI'21]. We answer this question positively. In fact, although $\mathbf{fvs}$ can be {\em arbitrarily smaller} than $\mathbf{fas}$, we attain a runtime of $\mathbf{fvs}^{\OO(\mathbf{fvs})}\cdot n^{\OO(1)}$. So, additionally, our work subsumes the best known algorithm for {\sc Tournament Fixing} with respect to $\mathbf{fas}$.

Tournament Fixing Parameterized by Feedback Vertex Set Number Is FPT

Discriminative pre-trained language models (PrLMs) learn to predict original texts from intentionally corrupted ones. Taking the former text as positive and the latter as negative samples, the PrLM can be trained effectively for contextualized representation. However, the training of such a type of PrLMs highly relies on the quality of the automatically constructed samples. Existing PrLMs simply treat all corrupted texts as equal negative without any examination, which actually lets the resulting model inevitably suffer from the false negative issue where training is carried out on pseudo-negative data and leads to less efficiency and less robustness in the resulting PrLMs. In this work, on the basis of defining the false negative issue in discriminative PrLMs that has been ignored for a long time, we design enhanced pre-training methods to counteract false negative predictions and encourage pre-training language models on true negatives by correcting the harmful gradient updates subject to false negative predictions. Experimental results on GLUE and SQuAD benchmarks show that our counter-false-negative pre-training methods indeed bring about better performance together with stronger robustness.

Language Model Pre-training on True Negatives

Multispectral object detection plays a vital role in safety-critical vision systems that require an around-the-clock operation and encounter dynamic real-world situations(e.g., self-driving cars and autonomous surveillance systems). Despite its crucial competence in safety-related applications, its security against physical attacks is severely understudied. We investigate the vulnerability of multispectral detectors against physical attacks by proposing a new physical method: Multispectral Invisible Coating. Utilizing transparent Low-e films, we realize a laminated visible-thermal physical attack by attaching Low-e films over a visible attack printing. Moreover, we apply our physical method to manufacture a Multispectral Invisible Suit that hides persons from the multiple view angles of Multispectral detectors. To simulate our attack under various surveillance scenes, we constructed a large-scale multispectral pedestrian dataset which we will release in public. Extensive experiments show that our proposed method effectively attacks the state-of-the-art multispectral detector both in the digital space and the physical world.

Multispectral Invisible Coating: Laminated Visible-Thermal Physical Attack against Multispectral Object Detectors using Transparent Low-e films

We study a fully online matching problem with stochastic arrivals and departures. In this model, each online arrival follows a known identical and independent distribution over a fixed set of agent types. Its sojourn time is unknown in advance and follows type-specific distributions with known expectations. The goal is to maximize the weighted reward from successful matches. To solve this problem, we first propose a linear program (LP)-based algorithm whose competitive ratio is lower bounded by 0.155 under mild conditions. We further achieve better ratios in some special cases. To demonstrate the challenges of the problem, we further establish several hardness results. In particular, we show that no online algorithm can achieve a competitive ratio better than 2/3 in this model and there is no LP-based algorithm (with respect to our proposed LP) with a competitive ratio better than 1/3. Finally, we demonstrate the effectiveness and efficiency of our algorithm  numerically. 

Fully Online Matching with Stochastic Arrivals and Departures

As a representative of public transportation, the fundamental issue of managing bike-sharing systems is bike flow prediction. Recent methods overemphasize the spatio-temporal correlations in the data, ignoring the effects of contextual conditions on the transportation system and the inter-regional time-varying causality. In addition, due to the disturbance of incomplete observations in the data, random contextual conditions lead to spurious correlations between data and features, making the prediction of the model ineffective in special scenarios. To overcome this issue, we propose a Spatio-temporal Neural Structure Causal Model(STNSCM) from the perspective of causality. First, we build a causal graph to describe the traffic prediction, and further analyze the causal relationship between the input data, contextual conditions, spatio-temporal states, and prediction results. Second, we propose to apply the frontdoor criterion to eliminate confounding biases in the feature extraction process. Finally, we propose a counterfactual representation reasoning module to extrapolate the spatio-temporal state under the factual scenario to future counterfactual scenarios to improve the prediction performance. Experiments on real-world datasets demonstrate the superior performance of our model, especially its resistance to fluctuations caused by the external environment. The source code and data will be released.

Spatio-temporal Neural Structural Causal Models for Bike Flow Prediction

Script is a kind of structured knowledge extracted from texts, which contains a sequence of events. Based on such knowledge, script event prediction aims to predict the subsequent event. To do so, two aspects should be considered for events, namely, event description (i.e., what the events should contain) and event encoding (i.e., how they should be encoded). Most existing methods describe an event by a verb together with a few core arguments (i.e., subject, object, and indirect object), which are not precise enough. In addition, existing event encoders are limited to a fixed number of arguments, which are not flexible enough to deal with extra information. Thus, in this paper, we propose the Rich Event Prediction (REP) framework for script event prediction. Fundamentally, it is based on the proposed rich event description, which enriches the existing ones with three kinds of important information, namely, the senses of verbs, extra semantic roles, and types of participants. REP contains an event extractor to extract such information from texts. Based on the extracted rich information, a predictor then selects the most probable subsequent event. The core component of the predictor is a transformer-based event encoder that integrates the above information flexibly. Experimental results on the widely used Gigaword Corpus show the effectiveness of the proposed framework.

Rich Event Modeling for Script Event Prediction

We study the problem of semantically annotating textual documents that are complex in the sense that the documents are long, feature rich, and domain specific. Due to their complexity, such annotation tasks require trained human workers, which are very expensive in both time and money. We propose CEMA, a method for deploying machine learning to assist humans in complex document annotation. CEMA employs a number of components to estimate the human cost of annotating each document and to select the set of documents to be annotated that strike the best balance between model accuracy and human cost. We conduct experiments on complex annotation tasks in which we compare CEMA against other document selection and annotation strategies. Our results show that CEMA is the most cost-efficient solution for those tasks.

Incremental-DETR: Incremental Few-Shot Object Detection via Self-Supervised Learning

Downloads

Next from AAAI 2023

Semi-Supervised Learning with Support Isolation by Small-Paced Self-Training

Similar lecture

FeedFormer: Revisiting Transformer Decoder for Efficient Semantic Segmentation