Existing open-vocabulary object detectors typically enlarge their vocabulary sizes by leveraging different forms of weak supervision. This helps generalize to novel objects at inference. Two popular forms of weak-supervision used in open-vocabulary detection (OVD) include pretrained CLIP model and image-level supervision. We note that both these modes of supervision are not optimally aligned for the detection task: CLIP is trained with image-text pairs and lacks precise localization of objects while the image-level supervision has been used with heuristics that do not accurately specify local object regions. In this work, we propose to address this problem by performing object-centric alignment of the language embeddings from the CLIP model. Furthermore, we visually ground the objects with only image-level supervision using a pseudo-labeling process that provides high-quality object proposals and helps expand the vocabulary during training. We establish a bridge between the above two object-alignment strategies via a novel weight transfer function that aggregates their complimentary strengths. In essence, the proposed model seeks to minimize the gap between object and image-centric representations in the OVD setting. On the COCO benchmark, our proposed approach achieves 36.6 AP50 on novel classes, an absolute 8.2 gain over the previous best performance. For LVIS, we surpass the state-of-the-art ViLD model by 5.0 mask AP for rare categories and 3.4 overall. Code: https://github.com/hanoonaR/object-centric-ovd.

We present the CUNI-Bergamot submission for the WMT22 General translation task. We compete in English$\rightarrow$Czech direction. Our submission further explores block backtranslation techniques. Compared to the previous work, we measure performance in terms of COMET score and named entities translation accuracy. We evaluate performance of MBR decoding compared to traditional mixed backtranslation training and we show a possible synergy when using both of the techniques simultaneously. The results show that both approaches are effective means of improving translation quality and they yield even better results when combined.

Designing faster optimization algorithms is of ever-growing interest. In recent years, learning to learn methods that learn how to optimize demonstrated very encouraging results. Current approaches usually do not effectively include the dynamics of the optimization process during training. They either omit it entirely or only implicitly assume the dynamics of an isolated parameter. In this paper, we show how to utilize the dynamic mode decomposition method for extracting informative features about optimization dynamics. By employing those features, we show that our learned optimizer generalizes much better to unseen optimization problems in short. The improved generalization is illustrated on multiple tasks where training the optimizer on one neural network generalizes to different architectures and distinct datasets.

Acquiring food items with a fork poses an immense challenge to a robot-assisted feeding system, due to the wide range of material properties and visual appearances present across food groups. Deformable foods necessitate different skewering strategies than firm ones, but inferring such characteristics for several previously unseen items on a plate remains nontrivial. Our key insight is to leverage visual and haptic observations during interaction with an item to rapidly and reactively plan skewering motions. We learn a generalizable, multimodal representation for a food item from raw sensory inputs which informs the optimal skewering strategy. Given this representation, we propose a zero-shot framework to sense visuo-haptic properties of a previously unseen item and reactively skewer it, all within a single interaction. Real-robot experiments with foods of varying levels of visual and textural diversity demonstrate that our multimodal policy outperforms baselines which do not exploit both visual and haptic cues or do not reactively plan. Across 6 plates of different food items, our proposed framework achieves 71% success over 69 skewering attempts total. Supplementary material, datasets, code, and videos are available on our website: https://sites.google.com/view/hapticvisualnet-corl22/home

The Modboat is a low-cost, underactuated, modular robot capable of surface swimming, docking to other modules, and undocking from them using only a single motor and two passive flippers. Undocking is achieved by causing intentional self-collision between the tails of neighboring modules in certain configurations; this becomes a challenge, however, when collective swimming as one connected component is desirable. Prior work has developed controllers that turn arbitrary configurations of docked Modboats into steerable vehicles, but they cannot counteract lateral forces and disturbances. In this work we present a centralized control strategy to create holonomic vehicles out of arbitrary configurations of docked Modboats using an iterative potential-field based search. We experimentally demonstrate that our controller performs well and can control surge and sway velocities and yaw angle simultaneously.

Figure 1 - Contributions of the Paper and Fair, Accessible, Interpretable and Reproducible (FAIR) AI adapted from (Halgamuge S., 2021) Artificial intelligence (AI) has seen an explosive growth over the last 20 years, largely through recent advances in machine learning (ML) - the data-centric branch of AI. A data-centric AI system consists of an AI model (a structure or architecture) and a method or learning algorithm that enables that model to derive usable information from data. Sometimes the data are exploratory, like the genomic data arriving from different parts of the world about constantly mutating viruses. To discover the presence of new variants or labels, we can feed an AI model with such uninterpreted data, so that researchers will be able to use this AI model to assign labels. Such AI models need unsupervised learning (UL) algorithms to extract information from unlabeled and uninterpreted data. We could also ask those researchers themselves to label data with appropriate variant labels, and feed both labels and genomic data to an AI model that can then use a supervised learning algorithm like deep learning (DL), so that it can serve as a predictor for known variants of the virus. If such an AI model of sufficient strength requires it to be large, deep and complex, we call it a deep neural network (DNN). Shallow neural networks, commonly referred to as Neural Networks (NNs) are data driven mathematical models consisting of about three layers of artificial neurons or nodes (several linear and nonlinear processing elements) which are interconnected through weighted connections.

The recent success and proliferation of machine learning and deep learning have provided powerful tools, which are also utilized for encrypted traffic analysis, classification, and threat detection in computer networks. These methods, neural networks in particular, are often complex and require a huge corpus of training data. Therefore, this paper focuses on collecting a large up-to-date dataset with almost 200 fine-grained service labels and 140 million network flows extended with packet-level metadata. The number of flows is three orders of magnitude higher than in other existing public labeled datasets of encrypted traffic. The number of service labels, which is important to make the problem hard and realistic, is four times higher than in the public dataset with the most class labels. The published dataset is intended as a benchmark for identifying services in encrypted traffic. Service identification can be further extended with the task of "rejecting" unknown services, i.e., the traffic not seen during the training phase. Neural networks offer superior performance for tackling this more challenging problem. To showcase the dataset's usefulness, we implemented a neural network with a multi-modal architecture, which is the state-of-the-art approach, and achieved 97.04% classification accuracy and detected 91.94% of unknown services with 5% false positive rate.

This paper focuses on analyzing and improving the commonsense ability of recent popular vision-language (VL) models. Despite the great success, we observe that existing VL-models still lack commonsense knowledge/reasoning ability (e.g., "Lemons are sour"), which is a vital component towards artificial general intelligence. Through our analysis, we find one important reason is that existing large-scale VL datasets do not contain much commonsense knowledge, which motivates us to improve the commonsense of VL-models from the data perspective. Rather than collecting a new VL training dataset, we propose a more scalable strategy, i.e., "Data Augmentation with kNowledge graph linearization for CommonsensE capability" (DANCE). It can be viewed as one type of data augmentation technique, which can inject commonsense knowledge into existing VL datasets on the fly during training. More specifically, we leverage the commonsense knowledge graph (e.g., ConceptNet) and create variants of text description in VL datasets via bidirectional sub-graph sequentialization. For better commonsense evaluation, we further propose the first retrieval-based commonsense diagnostic benchmark. By conducting extensive experiments on some representative VL-models, we demonstrate that our DANCE technique is able to significantly improve the commonsense ability while maintaining the performance on vanilla retrieval tasks. The code and data are available at https://github.com/pleaseconnectwifi/DANCE

Over the past years, fashion-related challenges have gained a lot of attention in the research community. Outfit generation and recommendation, i.e., the composition of a set of items of different types (e.g., tops, bottom, shoes, accessories) that go well together, are among the most challenging ones. That is because items have to be both compatible amongst each other and also personalized to match the taste of the customer. Recently there has been a plethora of work targeted at tackling these problems by adopting various techniques and algorithms from the machine learning literature. However, to date, there is no extensive comparison of the performance of the different algorithms for outfit generation and recommendation. In this paper, we close this gap by providing a broad evaluation and comparison of various algorithms, including both personalized and non-personalized approaches, using online, real-world user data from one of Europe's largest fashion stores. We present the adaptations we made to some of those models to make them suitable for personalized outfit generation. Moreover, we provide insights for models that have not yet been evaluated on this task, specifically, GPT, BERT and Seq-to-Seq LSTM.

Adversarial Imitation Learning (AIL) is a class of popular state-of-the-art Imitation Learning algorithms commonly used in robotics. In AIL, an artificial adversary's misclassification is used as a reward signal that is optimized by any standard Reinforcement Learning (RL) algorithm. Unlike most RL settings, the reward in AIL is $differentiable$ but current model-free RL algorithms do not make use of this property to train a policy. The reward is AIL is also shaped since it comes from an adversary. We leverage the differentiability property of the shaped AIL reward function and formulate a class of Actor Residual Critic (ARC) RL algorithms. ARC algorithms draw a parallel to the standard Actor-Critic (AC) algorithms in RL literature and uses a residual critic, $C$ function (instead of the standard $Q$ function) to approximate only the discounted future return (excluding the immediate reward). ARC algorithms have similar convergence properties as the standard AC algorithms with the additional advantage that the gradient through the immediate reward is exact. For the discrete (tabular) case with finite states, actions, and known dynamics, we prove that policy iteration with $C$ function converges to an optimal policy. In the continuous case with function approximation and unknown dynamics, we experimentally show that ARC aided AIL outperforms standard AIL in simulated continuous-control and real robotic manipulation tasks. ARC algorithms are simple to implement and can be incorporated into any existing AIL implementation with an AC algorithm. Video and link to code are available at: https://sites.google.com/view/actor-residual-critic.