在最近的几项研究中已经显示了过度参数化在实现卓越概括性能方面的好处，证明了在实践中使用较大模型的趋势。然而，在强大的学习背景下，神经网络大小的影响尚未得到很好的研究。在这项工作中，我们发现，在大量错误标记的示例的存在下，将网络大小的增加超出某个点可能是有害的。特别是，当标签噪声增加时，最初是单调或“双重下降”测试损失曲线（W.R.T.网络宽度）变成U形或双U形曲线，这表明某些模型具有中等大小的模型实现了最佳的概括。我们观察到，当通过随机修剪通过密度控制网络大小时，观察到相似的测试损失行为。我们还通过偏置变化分解和理论上表征标签噪声塑造方差项的方式来仔细研究现象。即使采用最新的鲁棒方法，也可以观察到测试损失的类似行为，这表明限制网络大小可以进一步提高现有方法。最后，我们从经验上检查网络大小对学习函数平稳性的影响，并发现最初的大小和平滑度之间的负相关性是由标签噪声翻转的。

最近已证明自我监督的对比学习（CL）非常有效地防止深网贴上嘈杂的标签。尽管取得了经验成功，但对对比度学习对增强鲁棒性的影响的理论理解非常有限。在这项工作中，我们严格地证明，通过对比度学习学到的表示矩阵可以通过：（i）与数据中每个子类相对应的一个突出的奇异值来增强鲁棒性，并显着较小的剩余奇异值； （ii）{{显着的单数矢量与每个子类的干净标签之间的一个很大的对齐。以上属性使对此类表示的线性层能够有效地学习干净的标签，而不会过度适应噪音。}我们进一步表明，通过对比度学习预先训练的深网的雅各比式的低级别结构使他们能够获得优越的最初的性能是在嘈杂的标签上进行微调时。最后，我们证明了对比度学习提供的最初鲁棒性使鲁棒训练方法能够在极端噪声水平下实现最先进的性能，例如平均27.18 \％\％和15.58 \％\％\％\％\％cifar-10上的提高和80 \％对称嘈杂标签的CIFAR-100，网络视频的准确性提高4.11 \％。

While the capabilities of autonomous systems have been steadily improving in recent years, these systems still struggle to rapidly explore previously unknown environments without the aid of GPS-assisted navigation. The DARPA Subterranean (SubT) Challenge aimed to fast track the development of autonomous exploration systems by evaluating their performance in real-world underground search-and-rescue scenarios. Subterranean environments present a plethora of challenges for robotic systems, such as limited communications, complex topology, visually-degraded sensing, and harsh terrain. The presented solution enables long-term autonomy with minimal human supervision by combining a powerful and independent single-agent autonomy stack, with higher level mission management operating over a flexible mesh network. The autonomy suite deployed on quadruped and wheeled robots was fully independent, freeing the human supervision to loosely supervise the mission and make high-impact strategic decisions. We also discuss lessons learned from fielding our system at the SubT Final Event, relating to vehicle versatility, system adaptability, and re-configurable communications.

Language models have become increasingly popular in recent years for tasks like information retrieval. As use-cases become oriented toward specific domains, fine-tuning becomes default for standard performance. To fine-tune these models for specific tasks and datasets, it is necessary to carefully tune the model's hyperparameters and training techniques. In this paper, we present an in-depth analysis of the performance of four transformer-based language models on the task of biomedical information retrieval. The models we consider are DeepMind's RETRO (7B parameters), GPT-J (6B parameters), GPT-3 (175B parameters), and BLOOM (176B parameters). We compare their performance on the basis of relevance, accuracy, and interpretability, using a large corpus of 480000 research papers on protein structure/function prediction as our dataset. Our findings suggest that smaller models, with <10B parameters and fine-tuned on domain-specific datasets, tend to outperform larger language models on highly specific questions in terms of accuracy, relevancy, and interpretability by a significant margin (+50% on average). However, larger models do provide generally better results on broader prompts.

Recent methods demonstrate that data augmentation using counterfactual knowledge can teach models the causal structure of a task, leading to robust and generalizable models. However, such counterfactual data often has a limited scale and diversity if crowdsourced and is computationally expensive to extend to new perturbation types if generated using supervised methods. To address this, we introduce a new framework called DISCO for automatically generating high-quality counterfactual data at scale. DISCO engineers prompts to generate phrasal perturbations with a large general language model. Then, a task-specific teacher model filters the generation to distill high-quality counterfactual data. We show that learning with this counterfactual data yields a comparatively small student model that is 6% (absolute) more robust and generalizes 5% better across distributions than baselines on various challenging evaluations. This model is also 15% more sensitive in differentiating original and counterfactual examples, on three evaluation sets written by human workers and via human-AI collaboration.

Multi-document summarization (MDS) has traditionally been studied assuming a set of ground-truth topic-related input documents is provided. In practice, the input document set is unlikely to be available a priori and would need to be retrieved based on an information need, a setting we call open-domain MDS. We experiment with current state-of-the-art retrieval and summarization models on several popular MDS datasets extended to the open-domain setting. We find that existing summarizers suffer large reductions in performance when applied as-is to this more realistic task, though training summarizers with retrieved inputs can reduce their sensitivity retrieval errors. To further probe these findings, we conduct perturbation experiments on summarizer inputs to study the impact of different types of document retrieval errors. Based on our results, we provide practical guidelines to help facilitate a shift to open-domain MDS. We release our code and experimental results alongside all data or model artifacts created during our investigation.

Language tasks involving character-level manipulations (e.g., spelling correction, many word games) are challenging for models based in subword tokenization. To address this, we adapt the interchange intervention training method of Geiger et al. (2021) to operate on type-level variables over characters. This allows us to encode robust, position-independent character-level information in the internal representations of subword-based models. We additionally introduce a suite of character-level tasks that systematically vary in their dependence on meaning and sequence-level context. While simple character-level tokenization approaches still perform best on purely form-based tasks like string reversal, our method is superior for more complex tasks that blend form, meaning, and context, such as spelling correction in context and word search games. Our approach also leads to subword-based models with human-intepretable internal representations of characters.

In data-driven systems, data exploration is imperative for making real-time decisions. However, big data is stored in massive databases that are difficult to retrieve. Approximate Query Processing (AQP) is a technique for providing approximate answers to aggregate queries based on a summary of the data (synopsis) that closely replicates the behavior of the actual data, which can be useful where an approximate answer to the queries would be acceptable in a fraction of the real execution time. In this paper, we discuss the use of Generative Adversarial Networks (GANs) for generating tabular data that can be employed in AQP for synopsis construction. We first discuss the challenges associated with constructing synopses in relational databases and then introduce solutions to those challenges. Following that, we organized statistical metrics to evaluate the quality of the generated synopses. We conclude that tabular data complexity makes it difficult for algorithms to understand relational database semantics during training, and improved versions of tabular GANs are capable of constructing synopses to revolutionize data-driven decision-making systems.

Trajectory-User Linking (TUL) is a relatively new mobility classification task in which anonymous trajectories are linked to the users who generated them. With applications ranging from personalized recommendations to criminal activity detection, TUL has received increasing attention over the past five years. While research has focused mainly on learning deep representations that capture complex spatio-temporal mobility patterns unique to individual users, we demonstrate that visit patterns are highly unique among users and thus simple heuristics applied directly to the raw data are sufficient to solve TUL. More specifically, we demonstrate that a single check-in per trajectory is enough to correctly predict the identity of the user up to 85% of the time. Moreover, by using a non-parametric classifier, we scale up TUL to over 100k users which is an increase over state-of-the-art by three orders of magnitude. Extensive empirical analysis on four real-world datasets (Brightkite, Foursquare, Gowalla and Weeplaces) compares our findings to state-of-the-art results, and more importantly validates our claim that TUL is easier than commonly believed.

The two popular datasets ScanRefer [16] and ReferIt3D [3] connect natural language to real-world 3D data. In this paper, we curate a large-scale and complementary dataset extending both the aforementioned ones by associating all objects mentioned in a referential sentence to their underlying instances inside a 3D scene. Specifically, our Scan Entities in 3D (ScanEnts3D) dataset provides explicit correspondences between 369k objects across 84k natural referential sentences, covering 705 real-world scenes. Crucially, we show that by incorporating intuitive losses that enable learning from this novel dataset, we can significantly improve the performance of several recently introduced neural listening architectures, including improving the SoTA in both the Nr3D and ScanRefer benchmarks by 4.3% and 5.0%, respectively. Moreover, we experiment with competitive baselines and recent methods for the task of language generation and show that, as with neural listeners, 3D neural speakers can also noticeably benefit by training with ScanEnts3D, including improving the SoTA by 13.2 CIDEr points on the Nr3D benchmark. Overall, our carefully conducted experimental studies strongly support the conclusion that, by learning on ScanEnts3D, commonly used visio-linguistic 3D architectures can become more efficient and interpretable in their generalization without needing to provide these newly collected annotations at test time. The project's webpage is https://scanents3d.github.io/ .