The goal of data attribution is to trace model predictions back to training data. Despite a long line of work towards this goal, existing approaches to data attribution tend to force users to choose between computational tractability and efficacy. That is, computationally tractable methods can struggle with accurately attributing model predictions in non-convex settings (e.g., in the context of deep neural networks), while methods that are effective in such regimes require training thousands of models, which makes them impractical for large models or datasets. In this work, we introduce TRAK (Tracing with the Randomly-projected After Kernel), a data attribution method that is both effective and computationally tractable for large-scale, differentiable models. In particular, by leveraging only a handful of trained models, TRAK can match the performance of attribution methods that require training thousands of models. We demonstrate the utility of TRAK across various modalities and scales: image classifiers trained on ImageNet, vision-language models (CLIP), and language models (BERT and mT5). We provide code for using TRAK (and reproducing our work) at https://github.com/MadryLab/trak .

Classic and deep generalized canonical correlation analysis (GCCA) algorithms seek low-dimensional common representations of data entities from multiple ``views'' (e.g., audio and image) using linear transformations and neural networks, respectively. When the views are acquired and stored at different computing agents (e.g., organizations and edge devices) and data sharing is undesired due to privacy or communication cost considerations, federated learning-based GCCA is well-motivated. In federated learning, the views are kept locally at the agents and only derived, limited information exchange with a central server is allowed. However, applying existing GCCA algorithms onto such federated learning settings may incur prohibitively high communication overhead. This work puts forth a communication-efficient federated learning framework for both linear and deep GCCA under the maximum variance (MAX-VAR) formulation. The overhead issue is addressed by aggressively compressing (via quantization) the exchanging information between the computing agents and a central controller. Compared to the unquantized version, our empirical study shows that the proposed algorithm enjoys a substantial reduction of communication overheads with virtually no loss in accuracy and convergence speed. Rigorous convergence analyses are also presented, which is a nontrivial effort. Generic federated optimization results do not cover the special problem structure of GCCA. Our result shows that the proposed algorithms for both linear and deep GCCA converge to critical points at a sublinear rate, even under heavy quantization and stochastic approximations. In addition, in the linear MAX-VAR case, the quantized algorithm approaches a global optimum in a geometric rate under reasonable conditions. Synthetic and real-data experiments are used to showcase the effectiveness of the proposed approach.

We focus on the problem of LiDAR point cloud based loop detection (or Finding) and closure (LDC) in a multi-agent setting. State-of-the-art (SOTA) techniques directly generate learned embeddings of a given point cloud, require large data transfers, and are not robust to wide variations in 6 Degrees-of-Freedom (DOF) viewpoint. Moreover, absence of strong priors in an unstructured point cloud leads to highly inaccurate LDC. In this original approach, we propose independent roll and pitch canonicalization of the point clouds using a common dominant ground plane. Discretization of the canonicalized point cloud along the axis perpendicular to the ground plane leads to an image similar to Digital Elevation Maps (DEMs), which exposes strong spatial priors in the scene. Our experiments show that LDC based on learnt embeddings of such DEMs is not only data efficient but also significantly more robust, and generalizable than the current SOTA. We report significant performance gain in terms of Average Precision for loop detection and absolute translation/rotation error for relative pose estimation (or loop closure) on Kitti, GPR and Oxford Robot Car over multiple SOTA LDC methods. Our encoder technique allows to compress the original point cloud by over 830 times. To further test the robustness of our technique we create and opensource a custom dataset called Lidar-UrbanFly Dataset (LUF) which consists of point clouds obtained from a LiDAR mounted on a quadrotor.

Keyword extraction is a crucial process in text mining. The extraction of keywords with respective contextual events in Twitter data is a big challenge. The challenging issues are mainly because of the informality in the language used. The use of misspelled words, acronyms, and ambiguous terms causes informality. The extraction of keywords with informal language in current systems is pattern based or event based. In this paper, contextual keywords are extracted using thematic events with the help of data association. The thematic context for events is identified using the uncertainty principle in the proposed system. The thematic contexts are weighed with the help of vectors called thematic context vectors which signifies the event as certain or uncertain. The system is tested on the Twitter COVID-19 dataset and proves to be effective. The system extracts event-specific thematic context vectors from the test dataset and ranks them. The extracted thematic context vectors are used for the clustering of contextual thematic vectors which improves the silhouette coefficient by 0.5% than state of art methods namely TF and TF-IDF. The thematic context vector can be used in other applications like Cyberbullying, sarcasm detection, figurative language detection, etc.

The growing number of cyber-attacks against Industrial Control Systems (ICS) in recent years has elevated security concerns due to the potential catastrophic impact. Considering the complex nature of ICS, detecting a cyber-attack in them is extremely challenging and requires advanced methods that can harness multiple data modalities. This research utilizes network and sensor modality data from ICS processed with a deep multi-modal cyber-attack detection model for ICS. Results using the Secure Water Treatment (SWaT) system show that the proposed model can outperform existing single modality models and recent works in the literature by achieving 0.99 precision, 0.98 recall, and 0.98 f-measure, which shows the effectiveness of using both modalities in a combined model for detecting cyber-attacks.

We develop a predictive inference procedure that combines conformal prediction (CP) with unconditional quantile regression (QR) -- a commonly used tool in econometrics that involves regressing the recentered influence function (RIF) of the quantile functional over input covariates. Unlike the more widely-known conditional QR, unconditional QR explicitly captures the impact of changes in covariate distribution on the quantiles of the marginal distribution of outcomes. Leveraging this property, our procedure issues adaptive predictive intervals with localized frequentist coverage guarantees. It operates by fitting a machine learning model for the RIFs using training data, and then applying the CP procedure for any test covariate with respect to a ``hypothetical'' covariate distribution localized around the new instance. Experiments show that our procedure is adaptive to heteroscedasticity, provides transparent coverage guarantees that are relevant to the test instance at hand, and performs competitively with existing methods in terms of efficiency.

Confusion matrices and derived metrics provide a comprehensive framework for the evaluation of model performance in machine learning. These are well-known and extensively employed in the supervised learning domain, particularly classification. Surprisingly, such a framework has not been fully explored in the context of clustering validation. Indeed, just recently such a gap has been bridged with the introduction of the Area Under the ROC Curve for Clustering (AUCC), an internal/relative Clustering Validation Index (CVI) that allows for clustering validation in real application scenarios. In this work we explore the Area Under Precision-Recall Curve (and related metrics) in the context of clustering validation. We show that these are not only appropriate as CVIs, but should also be preferred in the presence of cluster imbalance. We perform a comprehensive evaluation of proposed and state-of-art CVIs on real and simulated data sets. Our observations corroborate towards an unified validation framework for supervised and unsupervised learning, given that they are consistent with existing guidelines established for the evaluation of supervised learning models.

Our society is increasingly fond of computational tools. This phenomenon has greatly increased over the past decade following, among other factors, the emergence of a new Artificial Intelligence paradigm. Specifically, the coupling of two algorithmic techniques, Deep Neural Networks and Stochastic Gradient Descent, thrusted by an exponentially increasing computing capacity, has and is continuing to become a major asset in many modern technologies. However, as progress takes its course, some still wonder whether other methods could similarly or even more greatly benefit from these various hardware advances. In order to further this study, we delve in this thesis into Evolutionary Algorithms and their application to Dynamic Neural Networks, two techniques which despite enjoying many advantageous properties have yet to find their niche in contemporary Artificial Intelligence. We find that by elaborating new methods while exploiting strong computational resources, it becomes possible to develop strongly performing agents on a variety of benchmarks but also some other agents behaving very similarly to human subjects on the video game Shinobi III : Return of The Ninja Master, typical complex tasks previously out of reach for non-gradient-based optimization.

Today's computer vision models achieve human or near-human level performance across a wide variety of vision tasks. However, their architectures, data, and learning algorithms differ in numerous ways from those that give rise to human vision. In this paper, we investigate the factors that affect the alignment between the representations learned by neural networks and human mental representations inferred from behavioral responses. We find that model scale and architecture have essentially no effect on the alignment with human behavioral responses, whereas the training dataset and objective function both have a much larger impact. These findings are consistent across three datasets of human similarity judgments collected using two different tasks. Linear transformations of neural network representations learned from behavioral responses from one dataset substantially improve alignment with human similarity judgments on the other two datasets. In addition, we find that some human concepts such as food and animals are well-represented by neural networks whereas others such as royal or sports-related objects are not. Overall, although models trained on larger, more diverse datasets achieve better alignment with humans than models trained on ImageNet alone, our results indicate that scaling alone is unlikely to be sufficient to train neural networks with conceptual representations that match those used by humans. Representation learning is a fundamental part of modern computer vision systems, but the paradigm has its roots in cognitive science. When Rumelhart et al. (1986) developed backpropagation, their goal was to find a method that could learn representations of concepts that are distributed across neurons, similarly to the human brain. The discovery that representations learned by backpropagation could replicate nontrivial aspects of human concept learning was a key factor in its rise to popularity in the late 1980s (Sutherland, 1986; Ng & Hinton, 2017). A string of empirical successes has since shifted the primary focus of representation learning research away from its similarities to human cognition and toward practical applications. This shift has been fruitful. By some metrics, the best computer vision models now outperform the best individual humans on benchmarks such as ImageNet (Shankar et al., 2020; Beyer et al., 2020; Vasudevan et al., 2022). As computer vision systems become increasingly widely used outside of research, we would like to know if they see the world in the same way that humans do.

Machine learning (ML) based time series forecasting models often require and assume certain degrees of stationarity in the data when producing forecasts. However, in many real-world situations, the data distributions are not stationary and they can change over time while reducing the accuracy of the forecasting models, which in the ML literature is known as concept drift. Handling concept drift in forecasting is essential for many ML methods in use nowadays, however, the prior work only proposes methods to handle concept drift in the classification domain. To fill this gap, we explore concept drift handling methods in particular for Global Forecasting Models (GFM) which recently have gained popularity in the forecasting domain. We propose two new concept drift handling methods, namely: Error Contribution Weighting (ECW) and Gradient Descent Weighting (GDW), based on a continuous adaptive weighting concept. These methods use two forecasting models which are separately trained with the most recent series and all series, and finally, the weighted average of the forecasts provided by the two models are considered as the final forecasts. Using LightGBM as the underlying base learner, in our evaluation on three simulated datasets, the proposed models achieve significantly higher accuracy than a set of statistical benchmarks and LightGBM baselines across four evaluation metrics.