Topics Emerged in the Biomedical Field and Their Characteristics

This study aims to reveal what kind of topics emerged in the biomedical domain by retrospectively analyzing newly added MeSH (Medical Subject Headings) terms from 2001 to 2010 and how they have been used for indexing since their inclusion in the thesaurus. The goal is to investigate if the future trend of a new topic depends on what kind of topic it is without relying on external indicators such as growth, citation patterns, or word co-occurrences. This topic perspective complements the traditional publication perspective in studying emerging topics. Results show that topic characteristics, including topic category, clinical significance, and if a topic has any narrower terms at the time of inclusion, influence future popularity of a new MeSH. Four emergence trend patterns are identified, including emerged and sustained, emerged not sustained, emerged and fluctuated, and not yet emerged. Predictive models using topic characteristics for emerging topic prediction show promise. This suggests that the characteristics of topics and domain should be considered when predicting future emergence of research topics. This study bridges a gap in emerging topic prediction by offering a topic perspective and advocates for considering topic and domain characteristics as well as economic, medical, and environmental impact when studying emerging topics in the biomedical domain.

Reinforcement Learning-powered Semantic Communication via Semantic Similarity

We introduce a new semantic communication mechanism, whose key idea is to preserve the semantic information instead of strictly securing the bit-level precision. Starting by analyzing the defects of existing joint source channel coding (JSCC) methods, we show that the commonly used bit-level metrics are vulnerable of catching important semantic meaning and structures. To address this problem, we take advantage of learning from semantic similarity, instead of relying on conventional paired bit-level supervisions like cross entropy and bit error rate. However, to develop such a semantic communication system is indeed a nontrivial task, considering the nondifferentiability of most semantic metrics as well as the instability from noisy channels. To further resolve these issues, we put forward a reinforcement learning (RL)-based solution which allows us to simultaneously optimize any user-defined semantic measurement by using the policy gradient technique, and to interact with the surrounding noisy environment in a natural way. We have testified the proposed method in the challenging European-parliament dataset. Experiments on both AWGN and phase-invariant fading channel have confirmed the superiority of our method in revealing the semantic meanings, and better handling the channel noise especially in low-SNR situations. Apart from the experimental results, we further provide an indepth look at how the semantics model behaves, along with its superb generalization ability in real-life examples. As a brand new method in learning-based JSCC tasks, we also exemplify an RL-based image transmission paradigm, both to prove the generalization ability, and to leave this new topic for future discussion.

NMS-Loss: Learning with Non-Maximum Suppression for Crowded Pedestrian Detection

Non-Maximum Suppression (NMS) is essential for object detection and affects the evaluation results by incorporating False Positives (FP) and False Negatives (FN), especially in crowd occlusion scenes. In this paper, we raise the problem of weak connection between the training targets and the evaluation metrics caused by NMS and propose a novel NMS-Loss making the NMS procedure can be trained end-to-end without any additional network parameters. Our NMS-Loss punishes two cases when FP is not suppressed and FN is wrongly eliminated by NMS. Specifically, we propose a pull loss to pull predictions with the same target close to each other, and a push loss to push predictions with different targets away from each other. Experimental results show that with the help of NMS-Loss, our detector, namely NMS-Ped, achieves impressive results with Miss Rate of 5.92% on Caltech dataset and 10.08% on CityPersons dataset, which are both better than state-of-the-art competitors.

Bi-Granularity Contrastive Learning for Post-Training in Few-Shot Scene

The major paradigm of applying a pre-trained language model to downstream tasks is to fine-tune it on labeled task data, which often suffers instability and low performance when the labeled examples are scarce.~One way to alleviate this problem is to apply post-training on unlabeled task data before fine-tuning, adapting the pre-trained model to target domains by contrastive learning that considers either token-level or sequence-level similarity. Inspired by the success of sequence masking, we argue that both token-level and sequence-level similarities can be captured with a pair of masked sequences.~Therefore, we propose complementary random masking (CRM) to generate a pair of masked sequences from an input sequence for sequence-level contrastive learning and then develop contrastive masked language modeling (CMLM) for post-training to integrate both token-level and sequence-level contrastive learnings.~Empirical results show that CMLM surpasses several recent post-training methods in few-shot settings without the need for data augmentation.

BiTT: Bidirectional Tree Tagging for Joint Extraction of Overlapping Entities and Relations

Joint extraction refers to extracting triples, composed of entities and relations, simultaneously from the text with a single model. However, most existing methods fail to extract all triples accurately and efficiently from sentences with overlapping issue, i.e., the same entity is included in multiple triples. In this paper, we propose a novel scheme called Bidirectional Tree Tagging (BiTT) to label overlapping triples in text. In BiTT, the triples with the same relation category in a sentence are especially represented as two binary trees, each of which is converted into a word-level tags sequence to label each word. Based on BiTT scheme, we develop an end-to-end extraction framework to predict the BiTT tags and further extract triples efficiently. We adopt the Bi-LSTM and the BERT as the encoder in our framework respectively, and obtain promising results in public English as well as Chinese datasets.

Enhancing autonomy transparency: an option-centric rationale approach

While the advances in artificial intelligence and machine learning empower a new generation of autonomous systems for assisting human performance, one major concern arises from the human factors perspective: Humans have difficulty deciphering autonomy-generated solutions and increasingly perceive autonomy as a mysterious black box. The lack of transparency contributes to the lack of trust in autonomy and sub-optimal team performance. To enhance autonomy transparency, this study proposed an option-centric rationale display and evaluated its effectiveness. We developed a game Treasure Hunter wherein a human uncovers a map for treasures with the help from an intelligent assistant, and conducted a human-in-the-loop experiment with 34 participants. Results indicated that by conveying the intelligent assistants decision-making rationale via the option-centric rationale display, participants had higher trust in the system and calibrated their trust faster. Additionally, higher trust led to higher acceptance of recommendations from the intelligent assistant, and in turn higher task performance.

Weakly-Supervised Action Localization with Expectation-Maximization Multi-Instance Learning

Weakly-supervised action localization requires training a model to localize the action segments in the video given only video level action label. It can be solved under the Multiple Instance Learning (MIL) framework, where a bag (video) contains multiple instances (action segments). Since only the bags label is known, the main challenge is assigning which key instances within the bag to trigger the bags label. Most previous models use attention-based approaches applying attentions to generate the bags representation from instances, and then train it via the bags classification. These models, however, implicitly violate the MIL assumption that instances in negative bags should be uniformly negative. In this work, we explicitly model the key instances assignment as a hidden variable and adopt an Expectation-Maximization (EM) framework. We derive two pseudo-label generation schemes to model the E and M process and iteratively optimize the likelihood lower bound. We show that our EM-MIL approach more accurately models both the learning objective and the MIL assumptions. It achieves state-of-the-art performance on two standard benchmarks, THUMOS14 and ActivityNet1.2.