Soft targets combined with the cross-entropy loss have shown to improve generalization performance of deep neural networks on supervised classification tasks. The standard cross-entropy loss however assumes data to be categorically distributed, which may often not be the case in practice. In contrast, InfoNCE does not rely on such an explicit assumption but instead implicitly estimates the true conditional through negative sampling. Unfortunately, it cannot be combined with soft targets in its standard formulation, hindering its use in combination with sophisticated training strategies. In this paper, we address this limitation by proposing a principled loss function that is compatible with probabilistic targets. Our new soft target InfoNCE loss is conceptually simple, efficient to compute, and can be derived within the framework of noise contrastive estimation. Using a toy example, we demonstrate shortcomings of the categorical distribution assumption of cross-entropy, and discuss implications of sampling from soft distributions. We observe that soft target InfoNCE performs on par with strong soft target cross-entropy baselines and outperforms hard target NLL and InfoNCE losses on popular benchmarks, including ImageNet. Finally, we provide a simple implementation of our loss, geared towards supervised classification and fully compatible with deep classification model trained with cross-entropy.
相關內容
The estimation of directed couplings between the nodes of a network from indirect measurements is a central methodological challenge in scientific fields such as neuroscience, systems biology and economics. Unfortunately, the problem is generally ill-posed due to the possible presence of unknown delays in the measurements. In this paper, we offer a solution of this problem by using a variational Bayes framework, where the uncertainty over the delays is marginalized in order to obtain conservative coupling estimates. To overcome the well-known overconfidence of classical variational methods, we use a hybrid-VI scheme where the (possibly flat or multimodal) posterior over the measurement parameters is estimated using a forward KL loss while the (nearly convex) conditional posterior over the couplings is estimated using the highly scalable gradient-based VI. In our ground-truth experiments, we show that the network provides reliable and conservative estimates of the couplings, greatly outperforming similar methods such as regression DCM.
Recently, tremendous strides have been made to align the generation of Large Language Models (LLMs) with human values to mitigate toxic or unhelpful content. Leveraging Reinforcement Learning from Human Feedback (RLHF) proves effective and is widely adopted by researchers. However, implementing RLHF is complex, and its sensitivity to hyperparameters renders achieving stable performance and scalability challenging. Furthermore, prevailing approaches to preference alignment primarily concentrate on pairwise comparisons, with limited exploration into multi-response scenarios, thereby overlooking the potential richness within the candidate pool. For the above reasons, we propose a new approach: Listwise Reward Enhancement for Preference Alignment (LIRE), a gradient-based reward optimization approach that incorporates the offline rewards of multiple responses into a streamlined listwise framework, thus eliminating the need for online sampling during training. LIRE is straightforward to implement, requiring minimal parameter tuning, and seamlessly aligns with the pairwise paradigm while naturally extending to multi-response scenarios. Moreover, we introduce a self-enhancement algorithm aimed at iteratively refining the reward during training. Our experiments demonstrate that LIRE consistently outperforms existing methods across several benchmarks on dialogue and summarization tasks, with good transferability to out-of-distribution data, assessed using proxy reward models and human annotators.
Ongoing advances in microbiome profiling have allowed unprecedented insights into the molecular activities of microbial communities. This has fueled a strong scientific interest in understanding the critical role the microbiome plays in governing human health, by identifying microbial features associated with clinical outcomes of interest. Several aspects of microbiome data limit the applicability of existing variable selection approaches. In particular, microbiome data are high-dimensional, extremely sparse, and compositional. Importantly, many of the observed features, although categorized as different taxa, may play related functional roles. To address these challenges, we propose a novel compositional regression approach that leverages the data-adaptive clustering and variable selection properties of the spiked Dirichlet process to identify taxa that exhibit similar functional roles. Our proposed method, Bayesian Regression with Agglomerated Compositional Effects using a dirichLET process (BRACElet), enables the identification of a sparse set of features with shared impacts on the outcome, facilitating dimension reduction and model interpretation. We demonstrate that BRACElet outperforms existing approaches for microbiome variable selection through simulation studies and an application elucidating the impact of oral microbiome composition on insulin resistance.
Spiking neural networks play an important role in brain-like neuromorphic computations and in studying working mechanisms of neural circuits. One drawback of training a large scale spiking neural network is that updating all weights is quite expensive. Furthermore, after training, all information related to the computational task is hidden into the weight matrix, prohibiting us from a transparent understanding of circuit mechanisms. Therefore, in this work, we address these challenges by proposing a spiking mode-based training protocol, where the recurrent weight matrix is explained as a Hopfield-like multiplication of three matrices: input, output modes and a score matrix. The first advantage is that the weight is interpreted by input and output modes and their associated scores characterizing the importance of each decomposition term. The number of modes is thus adjustable, allowing more degrees of freedom for modeling the experimental data. This significantly reduces the training cost because of significantly reduced space complexity for learning. Training spiking networks is thus carried out in the mode-score space. The second advantage is that one can project the high dimensional neural activity (filtered spike train) in the state space onto the mode space which is typically of a low dimension, e.g., a few modes are sufficient to capture the shape of the underlying neural manifolds. We successfully apply our framework in two computational tasks -- digit classification and selective sensory integration tasks. Our method accelerate the training of spiking neural networks by a Hopfield-like decomposition, and moreover this training leads to low-dimensional attractor structures of high-dimensional neural dynamics.
Abstract reasoning problems pose challenges to the perception and cognition abilities of AI algorithms, demanding deeper pattern recognition and inductive reasoning beyond mere identification of explicit image features. In this study, we introduce PMoC, a probabilistic model tailored for the Bongard-Logo problem, achieving high reasoning accuracy through the construction of an conditional probabilistic model. Additionally, we have designed the Pose-Transformer, an enhanced Transformer-Encoder specifically crafted for complex abstract reasoning tasks, including Bongard-Logo, RAVEN, I-RAVEN, and PGM. Inspired by the pose matrix in capsule networks, Pose-Transformer strengthens the focus on positional relationships between local features when processing image data. When combined with PMoC, it can further enhance reasoning accuracy. Our Pose-Transformer effectively addresses reasoning difficulties associated with changes in the position of abstract entities, outperforming previous models on RAVEN's OIG, D3$\times$3 subsets, and the PGM dataset. Finally, considering the deployment difficulties arising from the large number of Pose-Transformer parameters, this paper presents a lightweight version, Straw-Pose-Transformer, which maintains performance while significantly reducing the parameter count. This study contributes to enhancing AI capabilities in abstract reasoning and cognitive pattern recognition.
This article aims to study efficient/trace optimal designs for crossover trials with multiple responses recorded from each subject in the time periods. A multivariate fixed effects model is proposed with direct and carryover effects corresponding to the multiple responses. The corresponding error dispersion matrix is chosen to be either of the proportional or the generalized Markov covariance type, permitting the existence of direct and cross-correlations within and between the multiple responses. The corresponding information matrices for direct effects under the two types of dispersions are used to determine efficient designs. The efficiency of orthogonal array designs of Type $I$ and strength $2$ is investigated for a wide choice of covariance functions, namely, Mat($0.5$), Mat($1.5$) and Mat($\infty$). To motivate these multivariate crossover designs, a gene expression dataset in a $3 \times 3$ framework is utilized.
Recurrent neural networks (RNNs) notoriously struggle to learn long-term memories, primarily due to vanishing and exploding gradients. The recent success of state-space models (SSMs), a subclass of RNNs, to overcome such difficulties challenges our theoretical understanding. In this paper, we delve into the optimization challenges of RNNs and discover that, as the memory of a network increases, changes in its parameters result in increasingly large output variations, making gradient-based learning highly sensitive, even without exploding gradients. Our analysis further reveals the importance of the element-wise recurrence design pattern combined with careful parametrizations in mitigating this effect. This feature is present in SSMs, as well as in other architectures, such as LSTMs. Overall, our insights provide a new explanation for some of the difficulties in gradient-based learning of RNNs and why some architectures perform better than others.
Diffusion models have emerged as effective distribution estimators in vision, language, and reinforcement learning, but their use as priors in downstream tasks poses an intractable posterior inference problem. This paper studies amortized sampling of the posterior over data, $\mathbf{x}\sim p^{\rm post}(\mathbf{x})\propto p(\mathbf{x})r(\mathbf{x})$, in a model that consists of a diffusion generative model prior $p(\mathbf{x})$ and a black-box constraint or likelihood function $r(\mathbf{x})$. We state and prove the asymptotic correctness of a data-free learning objective, relative trajectory balance, for training a diffusion model that samples from this posterior, a problem that existing methods solve only approximately or in restricted cases. Relative trajectory balance arises from the generative flow network perspective on diffusion models, which allows the use of deep reinforcement learning techniques to improve mode coverage. Experiments illustrate the broad potential of unbiased inference of arbitrary posteriors under diffusion priors: in vision (classifier guidance), language (infilling under a discrete diffusion LLM), and multimodal data (text-to-image generation). Beyond generative modeling, we apply relative trajectory balance to the problem of continuous control with a score-based behavior prior, achieving state-of-the-art results on benchmarks in offline reinforcement learning.
Hate detection has long been a challenging task for the NLP community. The task becomes complex in a code-mixed environment because the models must understand the context and the hate expressed through language alteration. Compared to the monolingual setup, we see very less work on code-mixed hate as large-scale annotated hate corpora are unavailable to make the study. To overcome this bottleneck, we propose using native language hate samples. We hypothesise that in the era of multilingual language models (MLMs), hate in code-mixed settings can be detected by majorly relying on the native language samples. Even though the NLP literature reports the effectiveness of MLMs on hate detection in many cross-lingual settings, their extensive evaluation in a code-mixed scenario is yet to be done. This paper attempts to fill this gap through rigorous empirical experiments. We considered the Hindi-English code-mixed setup as a case study as we have the linguistic expertise for the same. Some of the interesting observations we got are: (i) adding native hate samples in the code-mixed training set, even in small quantity, improved the performance of MLMs for code-mixed hate detection, (ii) MLMs trained with native samples alone observed to be detecting code-mixed hate to a large extent, (iii) The visualisation of attention scores revealed that, when native samples were included in training, MLMs could better focus on the hate emitting words in the code-mixed context, and (iv) finally, when hate is subjective or sarcastic, naively mixing native samples doesn't help much to detect code-mixed hate. We will release the data and code repository to reproduce the reported results.
Information theory, which describes the transmission of signals in the presence of noise, has enabled the development of reliable communication systems that underlie the modern world. Imaging systems can also be viewed as a form of communication, in which information about the object is "transmitted" through images. However, the application of information theory to imaging systems has been limited by the challenges of accounting for their physical constraints. Here, we introduce a framework that addresses these limitations by modeling the probabilistic relationship between objects and their measurements. Using this framework, we develop a method to estimate information using only a dataset of noisy measurements, without making any assumptions about the image formation process. We demonstrate that these estimates comprehensively quantify measurement quality across a diverse range of imaging systems and applications. Furthermore, we introduce Information-Driven Encoder Analysis Learning (IDEAL), a technique to optimize the design of imaging hardware for maximum information capture. This work provides new insights into the fundamental performance limits of imaging systems and offers powerful new tools for their analysis and design.
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