Pre-trained large language models demonstrate potential in extracting information from DNA sequences, yet adapting to a variety of tasks and data modalities remains a challenge. To address this, we propose DNAGPT, a generalized DNA pre-training model trained on over 200 billion base pairs from all mammals. By enhancing the classic GPT model with a binary classification task (DNA sequence order), a numerical regression task (guanine-cytosine content prediction), and a comprehensive token language, DNAGPT can handle versatile DNA analysis tasks while processing both sequence and numerical data. Our evaluation of genomic signal and region recognition, mRNA abundance regression, and artificial genomes generation tasks demonstrates DNAGPT's superior performance compared to existing models designed for specific downstream tasks, benefiting from pre-training using the newly designed model structure.
Being able to decorrelate a feature space from protected attributes is an area of active research and study in ethics, fairness, and also natural sciences. We introduce a novel decorrelation method using Convex Neural Optimal Transport Solvers (Cnots) that is able to decorrelate a continuous feature space against protected attributes with optimal transport. We demonstrate how well it performs in the context of jet classification in high energy physics, where classifier scores are desired to be decorrelated from the mass of a jet. The decorrelation achieved in binary classification approaches the levels achieved by the state-of-the-art using conditional normalising flows. When moving to multiclass outputs the optimal transport approach performs significantly better than the state-of-the-art, suggesting substantial gains at decorrelating multidimensional feature spaces.
The logistic regression model is one of the most popular data generation model in noisy binary classification problems. In this work, we study the sample complexity of estimating the parameters of the logistic regression model up to a given $\ell_2$ error, in terms of the dimension and the inverse temperature, with standard normal covariates. The inverse temperature controls the signal-to-noise ratio of the data generation process. While both generalization bounds and asymptotic performance of the maximum-likelihood estimator for logistic regression are well-studied, the non-asymptotic sample complexity that shows the dependence on error and the inverse temperature for parameter estimation is absent from previous analyses. We show that the sample complexity curve has two change-points in terms of the inverse temperature, clearly separating the low, moderate, and high temperature regimes.
Adversarial imitation learning (AIL) is a popular method that has recently achieved much success. However, the performance of AIL is still unsatisfactory on the more challenging tasks. We find that one of the major reasons is due to the low quality of AIL discriminator representation. Since the AIL discriminator is trained via binary classification that does not necessarily discriminate the policy from the expert in a meaningful way, the resulting reward might not be meaningful either. We propose a new method called Policy Contrastive Imitation Learning (PCIL) to resolve this issue. PCIL learns a contrastive representation space by anchoring on different policies and generates a smooth cosine-similarity-based reward. Our proposed representation learning objective can be viewed as a stronger version of the AIL objective and provide a more meaningful comparison between the agent and the policy. From a theoretical perspective, we show the validity of our method using the apprenticeship learning framework. Furthermore, our empirical evaluation on the DeepMind Control suite demonstrates that PCIL can achieve state-of-the-art performance. Finally, qualitative results suggest that PCIL builds a smoother and more meaningful representation space for imitation learning.
The development of reliable and fair diagnostic systems is often constrained by the scarcity of labeled data. To address this challenge, our work explores the feasibility of unsupervised domain adaptation (UDA) to integrate large external datasets for developing reliable classifiers. The adoption of UDA with multiple sources can simultaneously enrich the training set and bridge the domain gap between different skin lesion datasets, which vary due to distinct acquisition protocols. Particularly, UDA shows practical promise for improving diagnostic reliability when training with a custom skin lesion dataset, where only limited labeled data are available from the target domain. In this study, we investigate three UDA training schemes based on source data utilization: single-source, combined-source, and multi-source UDA. Our findings demonstrate the effectiveness of applying UDA on multiple sources for binary and multi-class classification. A strong correlation between test error and label shift in multi-class tasks has been observed in the experiment. Crucially, our study shows that UDA can effectively mitigate bias against minority groups and enhance fairness in diagnostic systems, while maintaining superior classification performance. This is achieved even without directly implementing fairness-focused techniques. This success is potentially attributed to the increased and well-adapted demographic information obtained from multiple sources.
We study universal rates for multiclass classification, establishing the optimal rates (up to log factors) for all hypothesis classes. This generalizes previous results on binary classification (Bousquet, Hanneke, Moran, van Handel, and Yehudayoff, 2021), and resolves an open question studied by Kalavasis, Velegkas, and Karbasi (2022) who handled the multiclass setting with a bounded number of class labels. In contrast, our result applies for any countable label space. Even for finite label space, our proofs provide a more precise bounds on the learning curves, as they do not depend on the number of labels. Specifically, we show that any class admits exponential rates if and only if it has no infinite Littlestone tree, and admits (near-)linear rates if and only if it has no infinite Daniely-Shalev-Shwartz-Littleston (DSL) tree, and otherwise requires arbitrarily slow rates. DSL trees are a new structure we define in this work, in which each node of the tree is given by a pseudo-cube of possible classifications of a given set of points. Pseudo-cubes are a structure, rooted in the work of Daniely and Shalev-Shwartz (2014), and recently shown by Brukhim, Carmon, Dinur, Moran, and Yehudayoff (2022) to characterize PAC learnability (i.e., uniform rates) for multiclass classification. We also resolve an open question of Kalavasis, Velegkas, and Karbasi (2022) regarding the equivalence of classes having infinite Graph-Littlestone (GL) trees versus infinite Natarajan-Littlestone (NL) trees, showing that they are indeed equivalent.
While ERM suffices to attain near-optimal generalization error in the stochastic learning setting, this is not known to be the case in the online learning setting, where algorithms for general concept classes rely on computationally inefficient oracles such as the Standard Optimal Algorithm (SOA). In this work, we propose an algorithm for online binary classification setting that relies solely on ERM oracle calls, and show that it has finite regret in the realizable setting and sublinearly growing regret in the agnostic setting. We bound the regret in terms of the Littlestone and threshold dimensions of the underlying concept class. We obtain similar results for nonparametric games, where the ERM oracle can be interpreted as a best response oracle, finding the best response of a player to a given history of play of the other players. In this setting, we provide learning algorithms that only rely on best response oracles and converge to approximate-minimax equilibria in two-player zero-sum games and approximate coarse correlated equilibria in multi-player general-sum games, as long as the game has a bounded fat-threshold dimension. Our algorithms apply to both binary-valued and real-valued games and can be viewed as providing justification for the wide use of double oracle and multiple oracle algorithms in the practice of solving large games.
Even though data annotation is extremely important for interpretability, research and development of artificial intelligence solutions, most research efforts such as active learning or few-shot learning focus on the sample efficiency problem. This paper studies the neglected complementary problem of getting annotated data given a predictor. For the simple binary classification setting, we present the spectrum ranging from optimal general solutions to practical efficient methods. The problem is framed as the full annotation of a binary classification dataset with the minimal number of yes/no questions when a predictor is available. For the case of general binary questions the solution is found in coding theory, where the optimal questioning strategy is given by the Huffman encoding of the possible labelings. However, this approach is computationally intractable even for small dataset sizes. We propose an alternative practical solution based on several heuristics and lookahead minimization of proxy cost functions. The proposed solution is analysed, compared with optimal solutions and evaluated on several synthetic and real-world datasets. On these datasets, the method allows a significant improvement ($23-86\%$) in annotation efficiency.
Can knowing where you are assist in perceiving objects in your surroundings, especially under adverse weather and lighting conditions? In this work we investigate whether a prior map can be leveraged to aid in the detection of dynamic objects in a scene without the need for a 3D map or pixel-level map-query correspondences. We contribute an algorithm which refines an initial set of candidate object detections and produces a refined subset of highly accurate detections using a prior map. We begin by using visual place recognition (VPR) to retrieve a reference map image for a given query image, then use a binary classification neural network that compares the query and mapping image regions to validate the query detection. Once our classification network is trained, on approximately 1000 query-map image pairs, it is able to improve the performance of vehicle detection when combined with an existing off-the-shelf vehicle detector. We demonstrate our approach using standard datasets across two cities (Oxford and Zurich) under different settings of train-test separation of map-query traverse pairs. We further emphasize the performance gains of our approach against alternative design choices and show that VPR suffices for the task, eliminating the need for precise ground truth localization.
This paper describes our participation in the MentalRiskES task at IberLEF 2023. The task involved predicting the likelihood of an individual experiencing depression based on their social media activity. The dataset consisted of conversations from 175 Telegram users, each labeled according to their evidence of suffering from the disorder. We used a combination of traditional machine learning and deep learning techniques to solve four predictive subtasks: binary classification, simple regression, multiclass classification, and multi-output regression. We approached this by training a model to solve the multi-output regression case and then transforming the predictions to work for the other three subtasks. We compare the performance of two modeling approaches: fine-tuning a BERT-based model directly for the task or using its embeddings as inputs to a linear regressor, with the latter yielding better results. The code to reproduce our results can be found at: https://github.com/simonsanvil/EarlyDepression-MentalRiskES
Research demonstrates learners engaging in the process of producing explanations to support their reasoning, can have a positive impact on learning. However, providing learners real-time explanatory feedback often presents challenges related to classification accuracy, particularly in domain-specific environments, containing situationally complex and nuanced responses. We present two approaches for supplying tutors real-time feedback within an online lesson on how to give students effective praise. This work-in-progress demonstrates considerable accuracy in binary classification for corrective feedback of effective, or effort-based (F1 score = 0.811), and ineffective, or outcome-based (F1 score = 0.350), praise responses. More notably, we introduce progress towards an enhanced approach of providing explanatory feedback using large language model-facilitated named entity recognition, which can provide tutors feedback, not only while engaging in lessons, but can potentially suggest real-time tutor moves. Future work involves leveraging large language models for data augmentation to improve accuracy, while also developing an explanatory feedback interface.
Inspired by Solomonoffs theory of inductive inference, we propose a prior based on circuit complexity. There are several advantages to this approach. First, it relies on a complexity measure that does not depend on the choice of UTM. There is one universal definition for Boolean circuits involving an universal operation such as nand with simple conversions to alternative definitions such as and, or, and not. Second, there is no analogue of the halting problem. The output value of a circuit can be calculated recursively by computer in time proportional to the number of gates, while a short program may run for a very long time. Our prior assumes that a Boolean function, or equivalently, Boolean string of fixed length, is generated by some Bayesian mixture of circuits. This model is appropriate for learning Boolean functions from partial information, a problem often encountered within machine learning as "binary classification." We argue that an inductive bias towards simple explanations as measured by circuit complexity is appropriate for this problem.
Support Vector Machine (SVM) has been one of the most successful machine learning techniques for binary classification problems. The key idea is to maximize the margin from the data to the hyperplane subject to correct classification on training samples. The commonly used hinge loss and its variations are sensitive to label noise, and unstable for resampling due to its unboundedness. This paper is concentrated on the kernel SVM with the $\ell_0$-norm hinge loss (referred as $\ell_0$-KSVM), which is a composite function of hinge loss and $\ell_0$-norm and then could overcome the difficulties mentioned above. In consideration of the nonconvexity and nonsmoothness of $\ell_0$-norm hinge loss, we first characterize the limiting subdifferential of the $\ell_0$-norm hinge loss and then derive the equivalent relationship among the proximal stationary point, the Karush-Kuhn-Tucker point, and the local optimal solution of $\ell_0$-KSVM. Secondly, we develop an ADMM algorithm for $\ell_0$-KSVM, and obtain that any limit point of the sequence generated by the proposed algorithm is a locally optimal solution. Lastly, some experiments on the synthetic and real datasets are illuminated to show that $\ell_0$-KSVM can achieve comparable accuracy compared with the standard KSVM while the former generally enjoys fewer support vectors.
Modern deep-learning systems are specialized to problem settings in which training occurs once and then never again, as opposed to continual-learning settings in which training occurs continually. If deep-learning systems are applied in a continual learning setting, then it is well known that they may fail to remember earlier examples. More fundamental, but less well known, is that they may also lose their ability to learn on new examples, a phenomenon called loss of plasticity. We provide direct demonstrations of loss of plasticity using the MNIST and ImageNet datasets repurposed for continual learning as sequences of tasks. In ImageNet, binary classification performance dropped from 89% accuracy on an early task down to 77%, about the level of a linear network, on the 2000th task. Loss of plasticity occurred with a wide range of deep network architectures, optimizers, activation functions, batch normalization, dropout, but was substantially eased by L2-regularization, particularly when combined with weight perturbation. Further, we introduce a new algorithm -- continual backpropagation -- which slightly modifies conventional backpropagation to reinitialize a small fraction of less-used units after each example and appears to maintain plasticity indefinitely.
While there has been a growing research interest in developing out-of-distribution (OOD) detection methods, there has been comparably little discussion around how these methods should be evaluated. Given their relevance for safe(r) AI, it is important to examine whether the basis for comparing OOD detection methods is consistent with practical needs. In this work, we take a closer look at the go-to metrics for evaluating OOD detection, and question the approach of exclusively reducing OOD detection to a binary classification task with little consideration for the detection threshold. We illustrate the limitations of current metrics (AUROC & its friends) and propose a new metric - Area Under the Threshold Curve (AUTC), which explicitly penalizes poor separation between ID and OOD samples. Scripts and data are available at https://github.com/glhr/beyond-auroc
One of the more significant obstacles in classification of skin cancer is the presence of artifacts. This paper investigates the effect of dark corner artifacts, which result from the use of dermoscopes, on the performance of a deep learning binary classification task. Previous research attempted to remove and inpaint dark corner artifacts, with the intention of creating an ideal condition for models. However, such research has been shown to be inconclusive due to lack of available datasets labelled with dark corner artifacts and detailed analysis and discussion. To address these issues, we label 10,250 skin lesion images from publicly available datasets and introduce a balanced dataset with an equal number of melanoma and non-melanoma cases. The training set comprises 6126 images without artifacts, and the testing set comprises 4124 images with dark corner artifacts. We conduct three experiments to provide new understanding on the effects of dark corner artifacts, including inpainted and synthetically generated examples, on a deep learning method. Our results suggest that introducing synthetic dark corner artifacts which have been superimposed onto the training set improved model performance, particularly in terms of the true negative rate. This indicates that deep learning learnt to ignore dark corner artifacts, rather than treating it as melanoma, when dark corner artifacts were introduced into the training set. Further, we propose a new approach to quantifying heatmaps indicating network focus using a root mean square measure of the brightness intensity in the different regions of the heatmaps. This paper provides a new guideline for skin lesions analysis with an emphasis on reproducibility.
Understanding the effect of a feature vector $x \in \mathbb{R}^d$ on the response value (label) $y \in \mathbb{R}$ is the cornerstone of many statistical learning problems. Ideally, it is desired to understand how a set of collected features combine together and influence the response value, but this problem is notoriously difficult, due to the high-dimensionality of data and limited number of labeled data points, among many others. In this work, we take a new perspective on this problem, and we study the question of assessing the difference of influence that the two given features have on the response value. We first propose a notion of closeness for the influence of features, and show that our definition recovers the familiar notion of the magnitude of coefficients in the parametric model. We then propose a novel method to test for the closeness of influence in general model-free supervised learning problems. Our proposed test can be used with finite number of samples with control on type I error rate, no matter the ground truth conditional law $\mathcal{L}(Y |X)$. We analyze the power of our test for two general learning problems i) linear regression, and ii) binary classification under mixture of Gaussian models, and show that under the proper choice of score function, an internal component of our test, with sufficient number of samples will achieve full statistical power. We evaluate our findings through extensive numerical simulations, specifically we adopt the datamodel framework (Ilyas, et al., 2022) for CIFAR-10 dataset to identify pairs of training samples with different influence on the trained model via optional black box training mechanisms.
Cyber attacks deceive machines into believing something that does not exist in the first place. However, there are some to which even humans fall prey. One such famous attack that attackers have used over the years to exploit the vulnerability of vision is known to be a Homoglyph attack. It employs a primary yet effective mechanism to create illegitimate domains that are hard to differentiate from legit ones. Moreover, as the difference is pretty indistinguishable for a user to notice, they cannot stop themselves from clicking on these homoglyph domain names. In many cases, that results in either information theft or malware attack on their systems. Existing approaches use simple, string-based comparison techniques applied in primary language-based tasks. Although they are impactful to some extent, they usually fail because they are not robust to different types of homoglyphs and are computationally not feasible because of their time requirement proportional to the string length. Similarly, neural network-based approaches are employed to determine real domain strings from fake ones. Nevertheless, the problem with both methods is that they require paired sequences of real and fake domain strings to work with, which is often not the case in the real world, as the attacker only sends the illegitimate or homoglyph domain to the vulnerable user. Therefore, existing approaches are not suitable for practical scenarios in the real world. In our work, we created GlyphNet, an image dataset that contains 4M domains, both real and homoglyphs. Additionally, we introduce a baseline method for a homoglyph attack detection system using an attention-based convolutional Neural Network. We show that our model can reach state-of-the-art accuracy in detecting homoglyph attacks with a 0.93 AUC on our dataset.
Unsupervised performance estimation, or evaluating how well models perform on unlabeled data is a difficult task. Recently, a method was proposed by Garg et al. [2022] which performs much better than previous methods. Their method relies on having a score function, satisfying certain properties, to map probability vectors outputted by the classifier to the reals, but it is an open problem which score function is best. We explore this problem by first showing that their method fundamentally relies on the ordering induced by this score function. Thus, under monotone transformations of score functions, their method yields the same estimate. Next, we show that in the binary classification setting, nearly all common score functions - the $L^\infty$ norm; the $L^2$ norm; negative entropy; and the $L^2$, $L^1$, and Jensen-Shannon distances to the uniform vector - all induce the same ordering over probability vectors. However, this does not hold for higher dimensional settings. We conduct numerous experiments on well-known NLP data sets and rigorously explore the performance of different score functions. We conclude that the $L^\infty$ norm is the most appropriate.
Algorithmic Fairness is an established field in machine learning that aims to reduce biases in data. Recent advances have proposed various methods to ensure fairness in a univariate environment, where the goal is to de-bias a single task. However, extending fairness to a multi-task setting, where more than one objective is optimised using a shared representation, remains underexplored. To bridge this gap, we develop a method that extends the definition of Strong Demographic Parity to multi-task learning using multi-marginal Wasserstein barycenters. Our approach provides a closed form solution for the optimal fair multi-task predictor including both regression and binary classification tasks. We develop a data-driven estimation procedure for the solution and run numerical experiments on both synthetic and real datasets. The empirical results highlight the practical value of our post-processing methodology in promoting fair decision-making.