Supervised finetuning (SFT) on instruction datasets has played a crucial role in achieving the remarkable zero-shot generalization capabilities observed in modern large language models (LLMs). However, the annotation efforts required to produce high quality responses for instructions are becoming prohibitively expensive, especially as the number of tasks spanned by instruction datasets continues to increase. Active learning is effective in identifying useful subsets of samples to annotate from an unlabeled pool, but its high computational cost remains a barrier to its widespread applicability in the context of LLMs. To mitigate the annotation cost of SFT and circumvent the computational bottlenecks of active learning, we propose using experimental design. Experimental design techniques select the most informative samples to label, and typically maximize some notion of uncertainty and/or diversity. In our work, we implement a framework that evaluates several existing and novel experimental design techniques and find that these methods consistently yield significant gains in label efficiency with little computational overhead. On generative tasks, our methods achieve the same generalization performance with only $50\%$ of annotation cost required by random sampling.
相關內容
A novel approach for achieving fast evasion in self-localized swarms of Unmanned Aerial Vehicles (UAVs) threatened by an intruding moving object is presented in this paper. Motivated by natural self-organizing systems, the presented approach of fast and collective evasion enables the UAV swarm to avoid dynamic objects (interferers) that are actively approaching the group. The main objective of the proposed technique is the fast and safe escape of the swarm from an interferer ~discovered in proximity. This method is inspired by the collective behavior of groups of certain animals, such as schools of fish or flocks of birds. These animals use the limited information of their sensing organs and decentralized control to achieve reliable and effective group motion. The system presented in this paper is intended to execute the safe coordination of UAV swarms with a large number of agents. Similar to natural swarms, this system propagates a fast shock of information about detected interferers throughout the group to achieve dynamic and collective evasion. The proposed system is fully decentralized using only onboard sensors to mutually localize swarm agents and interferers, similar to how animals accomplish this behavior. As a result, the communication structure between swarm agents is not overwhelmed by information about the state (position and velocity) of each individual and it is reliable to communication dropouts. The proposed system and theory were numerically evaluated and verified in real-world experiments.
Research in the area of secure multi-party computation with an unconventional method of using a physical deck of playing cards began in 1989 when den Boer proposed a protocol to compute the logical AND function using five cards. Since then, the area has gained interest from many researchers and several card-based protocols to compute various functions have been developed. In this paper, we propose a card-based protocol called the overwriting protocol that can securely compute the $k$-candidate $n$-variable equality function $f: \{0,1,\ldots ,k-1\}^n \rightarrow \{0,1\}$. We also apply the technique used in this protocol to compute other similar functions.
Designing effective game tutorials is crucial for a smooth learning curve for new players, especially in games with many rules and complex core mechanics. Evaluating the effectiveness of these tutorials usually requires multiple iterations with testers who have no prior knowledge of the game. Recent Vision-Language Models (VLMs) have demonstrated significant capabilities in understanding and interpreting visual content. VLMs can analyze images, provide detailed insights, and answer questions about their content. They can recognize objects, actions, and contexts in visual data, making them valuable tools for various applications, including automated game testing. In this work, we propose an automated game-testing solution to evaluate the quality of game tutorials. Our approach leverages VLMs to analyze frames from video game tutorials, answer relevant questions to simulate human perception, and provide feedback. This feedback is compared with expected results to identify confusing or problematic scenes and highlight potential errors for developers. In addition, we publish complete tutorial videos and annotated frames from different game versions used in our tests. This solution reduces the need for extensive manual testing, especially by speeding up and simplifying the initial development stages of the tutorial to improve the final game experience.
Automatic fingerprint recognition systems suffer from the threat of presentation attacks due to their wide range of deployment in areas including national borders and commercial applications. A presentation attack can be performed by creating a spoof of a user's fingerprint with or without their consent. This paper presents a dynamic ensemble of deep CNN and handcrafted features to detect presentation attacks in known-material and unknown-material protocols of the livness detection competition. The proposed presentation attack detection model, in this way, utilizes the capabilities of both deep CNN and handcrafted features techniques and exhibits better performance than their individual performances. We have validated our proposed method on benchmark databases from the Liveness Detection Competition in 2015, 2017, and 2019, yielding overall accuracy of 96.10\%, 96.49\%, and 94.99\% on them, respectively. The proposed method outperforms state-of-the-art methods in terms of classification accuracy.
An intrinsically causal approach to lifting factorization, called the Causal Complementation Algorithm, is developed for arbitrary two-channel perfect reconstruction FIR filter banks. This addresses an engineering shortcoming of the inherently noncausal strategy of Daubechies and Sweldens for factoring discrete wavelet transforms, which was based on the Extended Euclidean Algorithm for Laurent polynomials. The Causal Complementation Algorithm reproduces all lifting factorizations created by the causal version of the Euclidean Algorithm approach and generates additional causal factorizations, which are not obtainable via the causal Euclidean Algorithm, possessing degree-reducing properties that generalize those furnished by the Euclidean Algorithm. In lieu of the Euclidean Algorithm, the new approach employs Gaussian elimination in matrix polynomials using a slight generalization of polynomial long division. It is shown that certain polynomial degree-reducing conditions are both necessary and sufficient for a causal elementary matrix decomposition to be obtainable using the Causal Complementation Algorithm, yielding a formal definition of ``lifting factorization'' that was missing from the work of Daubechies and Sweldens.
Representation Misdirection for Unlearning (RMU), which steers model representation in the intermediate layer to a target random representation, is an effective method for large language model (LLM) unlearning. Despite its high performance, the underlying cause and explanation remain underexplored. In this paper, we first theoretically demonstrate that steering forget representations in the intermediate layer reduces token confidence, causing LLMs to generate wrong or nonsense responses. Second, we investigate how the coefficient influences the alignment of forget-sample representations with the random direction and hint at the optimal coefficient values for effective unlearning across different network layers. Third, we show that RMU unlearned models are robust against adversarial jailbreak attacks. Last, our empirical analysis shows that RMU is less effective when applied to the middle and later layers in LLMs. To resolve this drawback, we propose Adaptive RMU -- a simple yet effective alternative method that makes unlearning effective with most layers. Extensive experiments demonstrate that Adaptive RMU significantly improves the unlearning performance compared to prior art while incurring no additional computational cost.
Background. Fully automatic analysis of myocardial perfusion MRI datasets enables rapid and objective reporting of stress/rest studies in patients with suspected ischemic heart disease. Developing deep learning techniques that can analyze multi-center datasets despite limited training data and variations in software and hardware is an ongoing challenge. Methods. Datasets from 3 medical centers acquired at 3T (n = 150 subjects) were included: an internal dataset (inD; n = 95) and two external datasets (exDs; n = 55) used for evaluating the robustness of the trained deep neural network (DNN) models against differences in pulse sequence (exD-1) and scanner vendor (exD-2). A subset of inD (n = 85) was used for training/validation of a pool of DNNs for segmentation, all using the same spatiotemporal U-Net architecture and hyperparameters but with different parameter initializations. We employed a space-time sliding-patch analysis approach that automatically yields a pixel-wise "uncertainty map" as a byproduct of the segmentation process. In our approach, a given test case is segmented by all members of the DNN pool and the resulting uncertainty maps are leveraged to automatically select the "best" one among the pool of solutions. Results. The proposed DAUGS analysis approach performed similarly to the established approach on the internal dataset (p = n.s.) whereas it significantly outperformed on the external datasets (p < 0.005 for exD-1 and exD-2). Moreover, the number of image series with "failed" segmentation was significantly lower for the proposed vs. the established approach (4.3% vs. 17.1%, p < 0.0005). Conclusions. The proposed DAUGS analysis approach has the potential to improve the robustness of deep learning methods for segmentation of multi-center stress perfusion datasets with variations in the choice of pulse sequence, site location or scanner vendor.
The Mixture of Experts (MoE) approach is well-suited for multilingual and code-switching (CS) tasks due to its multi-expert architecture. This work introduces the DLG-MoE, a Dynamic Language Group-based MoE optimized for bilingual and CS scenarios. DLG-MoE operates based on a hierarchical routing mechanism. First, the language router explicitly models the language and dispatches the representations to the corresponding language expert groups. Subsequently, the unsupervised router within each language group implicitly models attributes beyond language, and coordinates expert routing and collaboration. The model achieves state-of-the-art (SOTA) performance while also having unparalleled flexibility. It supports different top-k inference and streaming capabilities, and can also prune the model parameters to obtain a monolingual sub-model. The Code will be released.
Multi-Label Text Classification (MLTC) is a practical yet challenging task that involves assigning multiple non-exclusive labels to each document. Previous studies primarily focus on capturing label correlations to assist label prediction by introducing special labeling schemes, designing specific model structures, or adding auxiliary tasks. Recently, the $k$ Nearest Neighbor ($k$NN) framework has shown promise by retrieving labeled samples as references to mine label co-occurrence information in the embedding space. However, two critical biases, namely embedding alignment bias and confidence estimation bias, are often overlooked, adversely affecting prediction performance. In this paper, we introduce a DEbiased Nearest Neighbors (DENN) framework for MLTC, specifically designed to mitigate these biases. To address embedding alignment bias, we propose a debiased contrastive learning strategy, enhancing neighbor consistency on label co-occurrence. For confidence estimation bias, we present a debiased confidence estimation strategy, improving the adaptive combination of predictions from $k$NN and inductive binary classifications. Extensive experiments conducted on four public benchmark datasets (i.e., AAPD, RCV1-V2, Amazon-531, and EUR-LEX57K) showcase the effectiveness of our proposed method. Besides, our method does not introduce any extra parameters.
High spectral dimensionality and the shortage of annotations make hyperspectral image (HSI) classification a challenging problem. Recent studies suggest that convolutional neural networks can learn discriminative spatial features, which play a paramount role in HSI interpretation. However, most of these methods ignore the distinctive spectral-spatial characteristic of hyperspectral data. In addition, a large amount of unlabeled data remains an unexploited gold mine for efficient data use. Therefore, we proposed an integration of generative adversarial networks (GANs) and probabilistic graphical models for HSI classification. Specifically, we used a spectral-spatial generator and a discriminator to identify land cover categories of hyperspectral cubes. Moreover, to take advantage of a large amount of unlabeled data, we adopted a conditional random field to refine the preliminary classification results generated by GANs. Experimental results obtained using two commonly studied datasets demonstrate that the proposed framework achieved encouraging classification accuracy using a small number of data for training.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191