The Ki-67 proliferation index is an essential biomarker that helps pathologists to diagnose and select appropriate treatments. However, automatic evaluation of Ki-67 is difficult due to nuclei overlapping and complex variations in their properties. This paper proposes an integrated pipeline for accurate automatic counting of Ki-67, where the impact of nuclei separation techniques is highlighted. First, semantic segmentation is performed by combining the Squeez and Excitation Resnet and Unet algorithms to extract nuclei from the background. The extracted nuclei are then divided into overlapped and non-overlapped regions based on eight geometric and statistical features. A marker-based Watershed algorithm is subsequently proposed and applied only to the overlapped regions to separate nuclei. Finally, deep features are extracted from each nucleus patch using Resnet18 and classified into positive or negative by a random forest classifier. The proposed pipeline's performance is validated on a dataset from the Department of Pathology at H\^opital Nord Franche-Comt\'e hospital.
Domain Adaptation (DA) is a typical problem in machine learning that aims to transfer the model trained on source domain to target domain with different distribution. Causal DA is a special case of DA that solves the problem from the view of causality. It embeds the probabilistic relationships in multiple domains in a larger causal structure network of a system and tries to find the causal source (or intervention) on the system as the reason of distribution drifts of the system states across domains. In this sense, causal DA is transformed as a causal discovery problem that finds invariant representation across domains through the conditional independence between the state variables and observable state of the system given interventions. Testing conditional independence is the corner stone of causal discovery. Recently, a copula entropy based conditional independence test was proposed with a rigorous theory and a non-parametric estimation method. In this paper, we first present a mathemetical model for causal DA problem and then propose a method for causal DA that finds the invariant representation across domains with the copula entropy based conditional independence test. The effectiveness of the method is verified on two simulated data. The power of the proposed method is then demonstrated on two real-world data: adult census income data and gait characteristics data.
This paper proposes an interaction reasoning network for modelling spatio-temporal relationships between hands and objects in video. The proposed interaction unit utilises a Transformer module to reason about each acting hand, and its spatio-temporal relation to the other hand as well as objects being interacted with. We show that modelling two-handed interactions are critical for action recognition in egocentric video, and demonstrate that by using positionally-encoded trajectories, the network can better recognise observed interactions. We evaluate our proposal on EPIC-KITCHENS and Something-Else datasets, with an ablation study.
Model quantization has emerged as an indispensable technique to accelerate deep learning inference. While researchers continue to push the frontier of quantization algorithms, existing quantization work is often unreproducible and undeployable. This is because researchers do not choose consistent training pipelines and ignore the requirements for hardware deployments. In this work, we propose Model Quantization Benchmark (MQBench), a first attempt to evaluate, analyze, and benchmark the reproducibility and deployability for model quantization algorithms. We choose multiple different platforms for real-world deployments, including CPU, GPU, ASIC, DSP, and evaluate extensive state-of-the-art quantization algorithms under a unified training pipeline. MQBench acts like a bridge to connect the algorithm and the hardware. We conduct a comprehensive analysis and find considerable intuitive or counter-intuitive insights. By aligning the training settings, we find existing algorithms have about the same performance on the conventional academic track. While for the hardware-deployable quantization, there is a huge accuracy gap which remains unsettled. Surprisingly, no existing algorithm wins every challenge in MQBench, and we hope this work could inspire future research directions.
Recently, large pre-trained neural language models have attained remarkable performance on many downstream natural language processing (NLP) applications via fine-tuning. In this paper, we target at how to further improve the token representations on the language models. We, therefore, propose a simple yet effective plug-and-play module, Sequential Attention Module (SAM), on the token embeddings learned from a pre-trained language model. Our proposed SAM consists of two main attention modules deployed sequentially: Feature-wise Attention Module (FAM) and Token-wise Attention Module (TAM). More specifically, FAM can effectively identify the importance of features at each dimension and promote the effect via dot-product on the original token embeddings for downstream NLP applications. Meanwhile, TAM can further re-weight the features at the token-wise level. Moreover, we propose an adaptive filter on FAM to prevent noise impact and increase information absorption. Finally, we conduct extensive experiments to demonstrate the advantages and properties of our proposed SAM. We first show how SAM plays a primary role in the champion solution of two subtasks of SemEval'21 Task 7. After that, we apply SAM on sentiment analysis and three popular NLP tasks and demonstrate that SAM consistently outperforms the state-of-the-art baselines.
This paper presents the PALI team's winning system for SemEval-2021 Task 2: Multilingual and Cross-lingual Word-in-Context Disambiguation. We fine-tune XLM-RoBERTa model to solve the task of word in context disambiguation, i.e., to determine whether the target word in the two contexts contains the same meaning or not. In the implementation, we first specifically design an input tag to emphasize the target word in the contexts. Second, we construct a new vector on the fine-tuned embeddings from XLM-RoBERTa and feed it to a fully-connected network to output the probability of whether the target word in the context has the same meaning or not. The new vector is attained by concatenating the embedding of the [CLS] token and the embeddings of the target word in the contexts. In training, we explore several tricks, such as the Ranger optimizer, data augmentation, and adversarial training, to improve the model prediction. Consequently, we attain first place in all four cross-lingual tasks.
Question answering from semi-structured tables can be seen as a semantic parsing task and is significant and practical for pushing the boundary of natural language understanding. Existing research mainly focuses on understanding contents from unstructured evidence, e.g., news, natural language sentences, and documents. The task of verification from structured evidence, such as tables, charts, and databases, is still less explored. This paper describes sattiy team's system in SemEval-2021 task 9: Statement Verification and Evidence Finding with Tables (SEM-TAB-FACT). This competition aims to verify statements and to find evidence from tables for scientific articles and to promote the proper interpretation of the surrounding article. In this paper, we exploited ensemble models of pre-trained language models over tables, TaPas and TaBERT, for Task A and adjust the result based on some rules extracted for Task B. Finally, in the leaderboard, we attain the F1 scores of 0.8496 and 0.7732 in Task A for the 2-way and 3-way evaluation, respectively, and the F1 score of 0.4856 in Task B.
This paper describes MagicPai's system for SemEval 2021 Task 7, HaHackathon: Detecting and Rating Humor and Offense. This task aims to detect whether the text is humorous and how humorous it is. There are four subtasks in the competition. In this paper, we mainly present our solution, a multi-task learning model based on adversarial examples, for task 1a and 1b. More specifically, we first vectorize the cleaned dataset and add the perturbation to obtain more robust embedding representations. We then correct the loss via the confidence level. Finally, we perform interactive joint learning on multiple tasks to capture the relationship between whether the text is humorous and how humorous it is. The final result shows the effectiveness of our system.
As the world ages, elderly care becomes a big concern of the society. To address the elderly's issues on dementia and fall risk, we have investigated smart cognitive and fall risk assessment with machine learning methodology based on the data collected from finger tapping test and Timed Up and Go (TUG) test. Meanwhile, we have discovered the associations between cognition and finger motion from finger tapping data and the association between fall risk and gait characteristics from TUG data. In this paper, we jointly analyze the finger tapping and gait characteristics data with copula entropy. We find that the associations between certain finger tapping characteristics (number of taps of both hand of bi-inphase and bi-untiphase) and TUG score, certain gait characteristics are relatively high. According to this finding, we propose to utilize this associations to improve the predictive models of automatic fall risk assessment we developed previously. Experimental results show that using the characteristics of both finger tapping and gait as inputs of the predictive models of predicting TUG score can considerably improve the prediction performance in terms of MAE compared with using only one type of characteristics.
This paper introduces EPIC-KITCHENS-100, the largest annotated egocentric dataset - 100 hrs, 20M frames, 90K actions - of wearable videos capturing long-term unscripted activities in 45 environments. This extends our previous dataset (EPIC-KITCHENS-55), released in 2018, resulting in more action segments (+128%), environments (+41%) and hours (+84%), using a novel annotation pipeline that allows denser and more complete annotations of fine-grained actions (54% more actions per minute). We evaluate the "test of time" - i.e. whether models trained on data collected in 2018 can generalise to new footage collected under the same hypotheses albeit "two years on". The dataset is aligned with 6 challenges: action recognition (full and weak supervision), detection, anticipation, retrieval (from captions), as well as unsupervised domain adaptation for action recognition. For each challenge, we define the task, provide baselines and evaluation metrics. Our dataset and challenge leaderboards will be made publicly available.