Solving strategic games whose action space is prohibitively large is a critical yet under-explored topic in economics, computer science and artificial intelligence. This paper proposes new learning algorithms in two-player zero-sum games where the number of pure strategies is huge or even infinite. Specifically, we combine no-regret analysis from online learning with double oracle methods from game theory. Our method -- \emph{Online Double Oracle (ODO)} -- achieves the regret bound of $\mathcal{O}(\sqrt{T k \log(k)})$ in self-play setting where $k$ is NOT the size of the game, but rather the size of \emph{effective strategy set} that is linearly dependent on the support size of the Nash equilibrium. On tens of different real-world games, including Leduc Poker that contains $3^{936}$ pure strategies, our methods outperform no-regret algorithms and double oracle methods by a large margin, both in convergence rate to Nash equilibrium and average payoff against strategic adversary.
Promoting behavioural diversity is critical for solving games with non-transitive dynamics where strategic cycles exist, and there is no consistent winner (e.g., Rock-Paper-Scissors). Yet, there is a lack of rigorous treatment for defining diversity and constructing diversity-aware learning dynamics. In this work, we offer a geometric interpretation of behavioural diversity in games and introduce a novel diversity metric based on \emph{determinantal point processes} (DPP). By incorporating the diversity metric into best-response dynamics, we develop \emph{diverse fictitious play} and \emph{diverse policy-space response oracle} for solving normal-form games and open-ended games. We prove the uniqueness of the diverse best response and the convergence of our algorithms on two-player games. Importantly, we show that maximising the DPP-based diversity metric guarantees to enlarge the \emph{gamescape} -- convex polytopes spanned by agents' mixtures of strategies. To validate our diversity-aware solvers, we test on tens of games that show strong non-transitivity. Results suggest that our methods achieve much lower exploitability than state-of-the-art solvers by finding effective and diverse strategies.
One of the leading single-channel speech separation (SS) models is based on a TasNet with a dual-path segmentation technique, where the size of each segment remains unchanged throughout all layers. In contrast, our key finding is that multi-granularity features are essential for enhancing contextual modeling and computational efficiency. We introduce a self-attentive network with a novel sandglass-shape, namely Sandglasset, which advances the state-of-the-art (SOTA) SS performance at significantly smaller model size and computational cost. Forward along each block inside Sandglasset, the temporal granularity of the features gradually becomes coarser until reaching half of the network blocks, and then successively turns finer towards the raw signal level. We also unfold that residual connections between features with the same granularity are critical for preserving information after passing through the bottleneck layer. Experiments show our Sandglasset with only 2.3M parameters has achieved the best results on two benchmark SS datasets -- WSJ0-2mix and WSJ0-3mix, where the SI-SNRi scores have been improved by absolute 0.8 dB and 2.4 dB, respectively, comparing to the prior SOTA results.
As the most economical and routine auxiliary examination in the diagnosis of root canal treatment, oral X-ray has been widely used by stomatologists. It is still challenging to segment the tooth root with a blurry boundary for the traditional image segmentation method. To this end, we propose a model for high-resolution segmentation based on polynomial curve fitting with landmark detection (HS-PCL). It is based on detecting multiple landmarks evenly distributed on the edge of the tooth root to fit a smooth polynomial curve as the segmentation of the tooth root, thereby solving the problem of fuzzy edge. In our model, a maximum number of the shortest distances algorithm (MNSDA) is proposed to automatically reduce the negative influence of the wrong landmarks which are detected incorrectly and deviate from the tooth root on the fitting result. Our numerical experiments demonstrate that the proposed approach not only reduces Hausdorff95 (HD95) by 33.9% and Average Surface Distance (ASD) by 42.1% compared with the state-of-the-art method, but it also achieves excellent results on the minute quantity of datasets, which greatly improves the feasibility of automatic root canal therapy evaluation by medical image computing.
Knowledge distillation~(KD) is an effective learning paradigm for improving the performance of lightweight student networks by utilizing additional supervision knowledge distilled from teacher networks. Most pioneering studies either learn from only a single teacher in their distillation learning methods, neglecting the potential that a student can learn from multiple teachers simultaneously, or simply treat each teacher to be equally important, unable to reveal the different importance of teachers for specific examples. To bridge this gap, we propose a novel adaptive multi-teacher multi-level knowledge distillation learning framework~(AMTML-KD), which consists two novel insights: (i) associating each teacher with a latent representation to adaptively learn instance-level teacher importance weights which are leveraged for acquiring integrated soft-targets~(high-level knowledge) and (ii) enabling the intermediate-level hints~(intermediate-level knowledge) to be gathered from multiple teachers by the proposed multi-group hint strategy. As such, a student model can learn multi-level knowledge from multiple teachers through AMTML-KD. Extensive results on publicly available datasets demonstrate the proposed learning framework ensures student to achieve improved performance than strong competitors.
Variable-density cellular structures can overcome connectivity and manufacturability issues of topologically-optimized, functionally graded structures, particularly when those structures are represented as discrete density maps. One na\"ive approach to creating variable-density cellular structures is simply replacing the discrete density map with an unselective type of unit cells having corresponding densities. However, doing so breaks the desired mechanical behavior, as equivalent density alone does not guarantee equivalent mechanical properties. Another approach uses homogenization methods to estimate each pre-defined unit cell's effective properties and remaps the unit cells following a scaling law. However, a scaling law merely mitigates the problem by performing an indirect and inaccurate mapping from the material property space to single-type unit cells. In contrast, we propose a deep generative model that resolves this problem by automatically learning an accurate mapping and generating diverse cellular unit cells conditioned on desired properties (i.e., Young's modulus and Poisson's ratio). We demonstrate our method via the use of implicit function-based unit cells and conditional generative adversarial networks. Results show that our method can 1) generate various unit cells that satisfy given material properties with high accuracy (relative error <5%), 2) create functionally graded cellular structures with high-quality interface connectivity (98.7% average overlap area at interfaces), and 3) improve the structural performance over the conventional topology-optimized variable-density structure (84.4% reduction in concentrated stress and extra 7% reduction in displacement).
We study the cocktail party problem and propose a novel attention network called Tune-In, abbreviated for training under negative environments with interference. It firstly learns two separate spaces of speaker-knowledge and speech-stimuli based on a shared feature space, where a new block structure is designed as the building block for all spaces, and then cooperatively solves different tasks. Between the two spaces, information is cast towards each other via a novel cross- and dual-attention mechanism, mimicking the bottom-up and top-down processes of a human's cocktail party effect. It turns out that substantially discriminative and generalizable speaker representations can be learnt in severely interfered conditions via our self-supervised training. The experimental results verify this seeming paradox. The learnt speaker embedding has superior discriminative power than a standard speaker verification method; meanwhile, Tune-In achieves remarkably better speech separation performances in terms of SI-SNRi and SDRi consistently in all test modes, and especially at lower memory and computational consumption, than state-of-the-art benchmark systems.
To extract robust deep representations from long sequential modeling of speech data, we propose a self-supervised learning approach, namely Contrastive Separative Coding (CSC). Our key finding is to learn such representations by separating the target signal from contrastive interfering signals. First, a multi-task separative encoder is built to extract shared separable and discriminative embedding; secondly, we propose a powerful cross-attention mechanism performed over speaker representations across various interfering conditions, allowing the model to focus on and globally aggregate the most critical information to answer the "query" (current bottom-up embedding) while paying less attention to interfering, noisy, or irrelevant parts; lastly, we form a new probabilistic contrastive loss which estimates and maximizes the mutual information between the representations and the global speaker vector. While most prior unsupervised methods have focused on predicting the future, neighboring, or missing samples, we take a different perspective of predicting the interfered samples. Moreover, our contrastive separative loss is free from negative sampling. The experiment demonstrates that our approach can learn useful representations achieving a strong speaker verification performance in adverse conditions.
Knowledge distillation (KD) has witnessed its powerful ability in learning compact models in deep learning field, but it is still limited in distilling localization information for object detection. Existing KD methods for object detection mainly focus on mimicking deep features between teacher model and student model, which not only is restricted by specific model architectures, but also cannot distill localization ambiguity. In this paper, we first propose localization distillation (LD) for object detection. In particular, our LD can be formulated as standard KD by adopting the general localization representation of bounding box. Our LD is very flexible, and is applicable to distill localization ambiguity for arbitrary architecture of teacher model and student model. Moreover, it is interesting to find that Self-LD, i.e., distilling teacher model itself, can further boost state-of-the-art performance. Second, we suggest a teacher assistant (TA) strategy to fill the possible gap between teacher model and student model, by which the distillation effectiveness can be guaranteed even the selected teacher model is not optimal. On benchmark datasets PASCAL VOC and MS COCO, our LD can consistently improve the performance for student detectors, and also boosts state-of-the-art detectors notably. Our source code and trained models are publicly available at https://github.com/HikariTJU/LD