Deep neural networks are susceptible to adversarially crafted, small and imperceptible changes in the natural inputs. The most effective defense mechanism against these examples is adversarial training which constructs adversarial examples during training by iterative maximization of loss. The model is then trained to minimize the loss on these constructed examples. This min-max optimization requires more data, larger capacity models, and additional computing resources. It also degrades the standard generalization performance of a model. Can we achieve robustness more efficiently? In this work, we explore this question from the perspective of knowledge transfer. First, we theoretically show the transferability of robustness from an adversarially trained teacher model to a student model with the help of mixup augmentation. Second, we propose a novel robustness transfer method called Mixup-Based Activated Channel Maps (MixACM) Transfer. MixACM transfers robustness from a robust teacher to a student by matching activated channel maps generated without expensive adversarial perturbations. Finally, extensive experiments on multiple datasets and different learning scenarios show our method can transfer robustness while also improving generalization on natural images.

On the other hand, Clearview Ai company has decided to maintain its innocence and defende that their business wasn't concerned with Australia and has no Australian druggies. The CEO of the New York grounded company, Hoan Ton-That, the practice was n't illegal and was done keeping all the laws and regulation in mind. He cleared that the facial images which were scrapped of the multitudinous social media platforms similar as LinkedIn, Facebook, Instagram, and others, were the bones available on open internet and didn't violate any law. H indeed expressed his disheartening and disappointment over the fact that he respects the country, its citizens and the officers who spent time and energy on the inquiry, but his technology was misinterpreted and devaluated.

My 21-year-old goddaughter, a second-year undergraduate, mentioned in passing that she watches video lectures offline at twice the normal speed. Struck by this, I asked some other students I know. Many now routinely accelerate their lectures when learning offline – often by 1.5 times, sometimes by more. Speed learning is not for everyone, but there are whole Reddit threads where students discuss how odd it will be to return to the lecture theatre. One contributor wrote: "Normal speed now sounds like drunk speed."

Now, it has raised $100 million to fuel its growth, a round of funding that values H2O.ai at $1.7 billion post-money ($1.6 billion pre-money). This is a Series E round, and it's being led by a strategic backer, the Commonwealth Bank of Australia (CBA), which has been a customer of the startup and will be using the backing to kick off a deeper partnership between the two to build new services. Others in the round include Goldman Sachs, Pivot Investment Partners, Crane Venture Partners and Celesta Capital. Further plans for the funding include building more products for H2O.ai as a whole, and hiring more talent to continue expanding the company's H2O AI Hybrid Cloud platform. This is not the first time that a customer has led a round as a strategic backer: in 2019, Goldman Sachs led the company's Series D of $72.5 million.

Several instance-based explainability methods for finding influential training examples for test-time decisions have been proposed recently, including Influence Functions, TraceIn, Representer Point Selection, Grad-Dot, and Grad-Cos. Typically these methods are evaluated using LOO influence (Cook's distance) as a gold standard, or using various heuristics. In this paper, we show that all of the above methods are unstable, i.e., extremely sensitive to initialization, ordering of the training data, and batch size. We suggest that this is a natural consequence of how in the literature, the influence of examples is assumed to be independent of model state and other examples -- and argue it is not. We show that LOO influence and heuristics are, as a result, poor metrics to measure the quality of instance-based explanations, and instead propose to evaluate such explanations by their ability to detect poisoning attacks. Further, we provide a simple, yet effective baseline to improve all of the above methods and show how it leads to very significant improvements on downstream tasks.

The first order derivative of a data density can be estimated efficiently by denoising score matching, and has become an important component in many applications, such as image generation and audio synthesis. Higher order derivatives provide additional local information about the data distribution and enable new applications. Although they can be estimated via automatic differentiation of a learned density model, this can amplify estimation errors and is expensive in high dimensional settings. To overcome these limitations, we propose a method to directly estimate high order derivatives (scores) of a data density from samples. We first show that denoising score matching can be interpreted as a particular case of Tweedie's formula. By leveraging Tweedie's formula on higher order moments, we generalize denoising score matching to estimate higher order derivatives. We demonstrate empirically that models trained with the proposed method can approximate second order derivatives more efficiently and accurately than via automatic differentiation. We show that our models can be used to quantify uncertainty in denoising and to improve the mixing speed of Langevin dynamics via Ozaki discretization for sampling synthetic data and natural images.

With the proliferation of knowledge graphs, modeling data with complex multirelational structure has gained increasing attention in the area of statistical relational learning. One of the most important goals of statistical relational learning is link prediction, i.e., predicting whether certain relations exist in the knowledge graph. A large number of models and algorithms have been proposed to perform link prediction, among which tensor factorization method has proven to achieve state-of-the-art performance in terms of computation efficiency and prediction accuracy. However, a common drawback of the existing tensor factorization models is that the missing relations and non-existing relations are treated in the same way, which results in a loss of information. To address this issue, we propose a binary tensor factorization model with probit link, which not only inherits the computation efficiency from the classic tensor factorization model but also accounts for the binary nature of relational data. Our proposed probit tensor factorization (PTF) model shows advantages in both the prediction accuracy and interpretability

Optimal transport distances (OT) have been widely used in recent work in Machine Learning as ways to compare probability distributions. These are costly to compute when the data lives in high dimension. Recent work aims specifically at reducing this cost by computing OT using low-rank projections of the data (seen as discrete measures) [Paty and Cuturi, 2019]. We extend this approach and show that one can approximate OT distances by using more general families of maps provided they are 1-Lipschitz. The best estimate is obtained by maximising OT over the given family. As OT calculations are done after mapping data to a lower dimensional space, our method scales well with the original data dimension. We demonstrate the idea with neural networks. We use Sinkhorn Divergences (SD) to approximate OT distances as they are differentiable and allow for gradientbased optimisation. We illustrate on synthetic data how our technique preserves accuracy and displays a low sensitivity of computational costs to the data dimension.

In real world, affecting the environment by a weak policy can be expensive or very risky, therefore hampers real world applications of reinforcement learning. Offline Reinforcement Learning (RL) can learn policies from a given dataset without interacting with the environment. However, the dataset is the only source of information for an Offline RL algorithm and determines the performance of the learned policy. We still lack studies on how dataset characteristics influence different Offline RL algorithms. Therefore, we conducted a comprehensive empirical analysis of how dataset characteristics effect the performance of Offline RL algorithms for discrete action environments. A dataset is characterized by two metrics: (1) the average dataset return measured by the Trajectory Quality (TQ) and (2) the coverage measured by the State-Action Coverage (SACo). We found that variants of the off-policy Deep Q-Network family require datasets with high SACo to perform well. Algorithms that constrain the learned policy towards the given dataset perform well for datasets with high TQ or SACo. For datasets with high TQ, Behavior Cloning outperforms or performs similarly to the best Offline RL algorithms.

We propose a model-free reinforcement learning method for controlling mixed autonomy traffic in simulated traffic networks with through-traffic-only two-way and four-way intersections. Our method utilizes multi-agent policy decomposition which allows decentralized control based on local observations for an arbitrary number of controlled vehicles. We demonstrate that, even without reward shaping, reinforcement learning learns to coordinate the vehicles to exhibit traffic signal-like behaviors, achieving near-optimal throughput with 33-50% controlled vehicles. With the help of multi-task learning and transfer learning, we show that this behavior generalizes across inflow rates and size of the traffic network. Our code, models, and videos of results are available at https://github.com/ZhongxiaYan/mixed_autonomy_intersections.