As an aspiring data scientist, you must have heard the advice "do data science projects" over a thousand times. Not only are data science projects a great learning experience, they also help you stand out from the crowd of data science enthusiasts looking to break into the field. In this article, I am going to walk you through the projects that are must-haves on your resume. I will also provide you with sample datasets to experiment with for each project, along with associated tutorials that will help you complete the project. Data collection and pre-processing is one of the most important skills to have as a data scientist.

Existing work on understanding deep learning often employs measures that compress all data-dependent information into a few numbers. In this work, we adopt a perspective based on the role of individual examples. We introduce a measure of the computational difficulty of making a prediction for a given input: the (effective) prediction depth. Our extensive investigation reveals surprising yet simple relationships between the prediction depth of a given input and the model's uncertainty, confidence, accuracy and speed of learning for that data point. We further categorize difficult examples into three interpretable groups, demonstrate how these groups are processed differently inside deep models and showcase how this understanding allows us to improve prediction accuracy. Insights from our study lead to a coherent view of a number of separately reported phenomena in the literature: early layers generalize while later layers memorize; early layers converge faster and networks learn easy data and simple functions first.

Imitation learning (IL) algorithms have shown promising results for robots to learn skills from expert demonstrations. However, for versatile robots nowadays that need to learn diverse tasks, providing and learning the multi-task demonstrations all at once are both difficult. To solve this problem, in this work we study how to realize continual imitation learning ability that empowers robots to continually learn new tasks one by one, thus reducing the burden of multi-task IL and accelerating the process of new task learning at the same time. We propose a novel trajectory generation model that employs both a generative adversarial network and a dynamics prediction model to generate pseudo trajectories from all learned tasks in the new task learning process to achieve continual imitation learning ability. Our experiments on both simulation and real world manipulation tasks demonstrate the effectiveness of our method.

This paper presents an end-to-end semi-supervised object detection approach, in contrast to previous more complex multi-stage methods. The end-to-end training gradually improves pseudo label qualities during the curriculum, and the more and more accurate pseudo labels in turn benefit object detection training. We also propose two simple yet effective techniques within this framework: a soft teacher mechanism where the classification loss of each unlabeled bounding box is weighed by the classification score produced by the teacher network; a box jittering approach to select reliable pseudo boxes for the learning of box regression. On COCO benchmark, the proposed approach outperforms previous methods by a large margin under various labeling ratios, i.e. 1\%, 5\% and 10\%. Moreover, our approach proves to perform also well when the amount of labeled data is relatively large. For example, it can improve a 40.9 mAP baseline detector trained using the full COCO training set by +3.6 mAP, reaching 44.5 mAP, by leveraging the 123K unlabeled images of COCO. On the state-of-the-art Swin Transformer-based object detector (58.9 mAP on test-dev), it can still significantly improve the detection accuracy by +1.5 mAP, reaching 60.4 mAP, and improve the instance segmentation accuracy by +1.2 mAP, reaching 52.4 mAP, pushing the new state-of-the-art.

While machine learning algorithms excel at many challenging visual tasks, it is unclear that they can make predictions about commonplace real world physical events. Here, we present a visual and physical prediction benchmark that precisely measures this capability. In realistically simulating a wide variety of physical phenomena -- rigid and soft-body collisions, stable multi-object configurations, rolling and sliding, projectile motion -- our dataset presents a more comprehensive challenge than existing benchmarks. Moreover, we have collected human responses for our stimuli so that model predictions can be directly compared to human judgments. We compare an array of algorithms -- varying in their architecture, learning objective, input-output structure, and training data -- on their ability to make diverse physical predictions. We find that graph neural networks with access to the physical state best capture human behavior, whereas among models that receive only visual input, those with object-centric representations or pretraining do best but fall far short of human accuracy. This suggests that extracting physically meaningful representations of scenes is the main bottleneck to achieving human-like visual prediction. We thus demonstrate how our benchmark can identify areas for improvement and measure progress on this key aspect of physical understanding.

Multiple views of data, both naturally acquired (e.g., image and audio) and artificially produced (e.g., via adding different noise to data samples), have proven useful in enhancing representation learning. Natural views are often handled by multiview analysis tools, e.g., (deep) canonical correlation analysis [(D)CCA], while the artificial ones are frequently used in self-supervised learning (SSL) paradigms, e.g., SimCLR and Barlow Twins. Both types of approaches often involve learning neural feature extractors such that the embeddings of data exhibit high cross-view correlations. Although intuitive, the effectiveness of correlation-based neural embedding is only empirically validated. This work puts forth a theory-backed framework for unsupervised multiview learning. Our development starts with proposing a multiview model, where each view is a nonlinear mixture of shared and private components. Consequently, the learning problem boils down to shared/private component identification and disentanglement. Under this model, latent correlation maximization is shown to guarantee the extraction of the shared components across views (up to certain ambiguities). In addition, the private information in each view can be provably disentangled from the shared using proper regularization design. The method is tested on a series of tasks, e.g., downstream clustering, which all show promising performance. Our development also provides a unifying perspective for understanding various DCCA and SSL schemes.

Federated learning involves training machine learning models over devices or data silos, such as edge processors or data warehouses, while keeping the data local. Training in heterogeneous and potentially massive networks introduces bias into the system, which is originated from the non-IID data and the low participation rate in reality. In this paper, we propose Elastic Federated Learning (EFL), an unbiased algorithm to tackle the heterogeneity in the system, which makes the most informative parameters less volatile during training, and utilizes the incomplete local updates. It is an efficient and effective algorithm that compresses both upstream and downstream communications. Theoretically, the algorithm has convergence guarantee when training on the non-IID data at the low participation rate. Empirical experiments corroborate the competitive performance of EFL framework on the robustness and the efficiency.

The bulk of the algorithmic fairness literature deals with group fairness along the lines of demographic parity [9] or equal opportunity We study a novel problem of fairness in ranking aimed at minimizing [16]: this is typically expressed by means of some fairness the amount of individual unfairness introduced when enforcing constraint requiring that the top-positions (for any) in the ranking group-fairness constraints. Our proposal is rooted in the contain enough elements from some groups that are protected distributional maxmin fairness theory, which uses randomization from discrimination based on sex, race, age, etc. In fact, [6] shows to maximize the expected satisfaction of the worst-off individuals.

The joint use of multiple imaging modalities for medical image segmentation has been widely studied in recent years. The fusion of information from different modalities has demonstrated to improve the segmentation accuracy, with respect to mono-modal segmentations, in several applications. However, acquiring multiple modalities is usually not possible in a clinical setting due to a limited number of physicians and scanners, and to limit costs and scan time. Most of the time, only one modality is acquired. In this paper, we propose KD-Net, a framework to transfer knowledge from a trained multi-modal network (teacher) to a mono-modal one (student). The proposed method is an adaptation of the generalized distillation framework where the student network is trained on a subset (1 modality) of the teacher's inputs (n modalities). We illustrate the effectiveness of the proposed framework in brain tumor segmentation with the BraTS 2018 dataset. Using different architectures, we show that the student network effectively learns from the teacher and always outperforms the baseline mono-modal network in terms of segmentation accuracy.

This is the 12th "Future of the Internet" canvassing Pew Research Center and Elon University's Imagining the Internet Center have conducted together to get expert views about important digital issues. In this case, the questions focused on the prospects for ethical artificial intelligence (AI) by the year 2030. This is a nonscientific canvassing based on a nonrandom sample; this broad array of opinions about where current trends may lead in the next decade represents only the points of view of the individuals who responded to the queries. Pew Research and Elon's Imagining the Internet Center built a database of experts to canvass from a wide range of fields, choosing to invite people from several sectors, including professionals and policy people based in government bodies, nonprofits and foundations, technology businesses, think tanks and in networks of interested academics and technology innovators. The predictions reported here came in response to a set of questions in an online canvassing conducted between June 30 and July 27, 2020. In all, 602 technology innovators and developers, business and policy leaders, researchers and activists responded to at least one of the questions covered in this report. More on the methodology underlying this canvassing and the participants can be found in the final section. Artificial intelligence systems "understand" and shape a lot of what happens in people's lives. AI applications "speak" to people and answer questions when the name of a digital voice assistant is called out. They run the chatbots that handle customer-service issues people have with companies. They help diagnose cancer and other medical conditions. They scour the use of credit cards for signs of fraud, and they determine who could be a credit risk. They help people drive from point A to point B and update traffic information to shorten travel times. They are the operating system of driverless vehicles. They sift applications to make recommendations about job candidates. They determine the material that is offered up in people's newsfeeds and video choices. They recognize people's faces, translate languages and suggest how to complete people's sentences or search queries. They can "read" people's emotions. They beat them at sophisticated games.