Haidong Zhang - Academia.edu (original) (raw)

Papers by Haidong Zhang

Cornell University - arXiv, May 4, 2021

Distributional text clustering delivers semantically informative representations and captures the... more Distributional text clustering delivers semantically informative representations and captures the relevance between each word and semantic clustering centroids. We extend the neural text clustering approach to text classification tasks by inducing cluster centers via a latent variable model and interacting with distributional word embeddings, to enrich the representation of tokens and measure the relatedness between tokens and each learnable cluster centroid. The proposed method jointly learns word clustering centroids and clustering-token alignments, achieving the state of the art results on multiple benchmark datasets and proving that the proposed cluster-token alignment mechanism is indeed favorable to text classification. Notably, our qualitative analysis has conspicuously illustrated that text representations learned by the proposed model are in accord well with our intuition. * Equal contribution. Preprint. Under review.

Findings of the Association for Computational Linguistics: EMNLP 2021, 2021

Generative Adversarial Networks (GANs) have achieved great success in image synthesis, but have p... more Generative Adversarial Networks (GANs) have achieved great success in image synthesis, but have proven to be difficult to generate natural language. Challenges arise from the uninformative learning signals passed from the discriminator. In other words, the poor learning signals limit the learning capacity for generating languages with rich structures and semantics. In this paper, we propose to adopt the counter-contrastive learning (CCL) method to support the generator's training in language GANs. In contrast to standard GANs that adopt a simple binary classifier to discriminate whether a sample is real or fake, we employ a counter-contrastive learning signal that advances the training of language synthesizers by (1) pulling the language representations of generated and real samples together and (2) pushing apart representations of real samples to compete with the discriminator and thus prevent the discriminator from being overtrained. We evaluate our method on both synthetic and real benchmarks and yield competitive performance compared to previous GANs for adversarial sequence generation.

Journal of Cleaner Production, 2021

Abstract Flotation has a great potential in the separation of complicated poly-mixtures to obtain... more Abstract Flotation has a great potential in the separation of complicated poly-mixtures to obtain qualified resins for recycling, but it remains a technological challenge for selective wetting between components. This study provided a combined flotation process with the prepositive plasma modification, separating hazardous polyvinyl chloride (PVC) from recyclable high-density polyethylene (HDPE) and polyethylene terephthalate (PET). The plasma process was characterized by surface morphology, molecular weight, contact angle, surface free energy and spectrum. The results suggested that amorphous low molecular weight oxidic substances can be selectively introduced on these polymers through the hydrogen extraction and Norrish І type radical pathways in plasma zone, leading to a desired surface wetting. Under the optimal unit activating energy of 12.0 kJ/m2 (PVC-HDPE) and 15.0 kJ/m2 (PVC-PET), separated PVC can reach above 93% of both recovery and purity. Benefiting from the green and mild operation in plasma unit, this study paves a new way to effectively separate waste plastics as an alternative to traditional surface modification with chemical reagents.

Proceedings of the Third Workshop on Multimodal Artificial Intelligence, 2021

Emotion recognition in conversation has received considerable attention recently because of its p... more Emotion recognition in conversation has received considerable attention recently because of its practical industrial applications. Existing methods tend to overlook the immediate mutual interaction between different speakers in the speaker-utterance level, or apply single speaker-agnostic RNN for utterances from different speakers. We propose COIN, a conversational interactive model to mitigate this problem by applying state mutual interaction within history contexts. In addition, we introduce a stacked global interaction module to capture the contextual and inter-dependency representation in a hierarchical manner. To improve the robustness and generalization during training, we generate adversarial examples by applying the minor perturbations on multimodal feature inputs, unveiling the benefits of adversarial examples for emotion detection. The proposed model empirically achieves the current state-of-the-art results on the IEMO-CAP benchmark dataset.

Cornell University - arXiv, May 4, 2021

Distributional text clustering delivers semantically informative representations and captures the... more Distributional text clustering delivers semantically informative representations and captures the relevance between each word and semantic clustering centroids. We extend the neural text clustering approach to text classification tasks by inducing cluster centers via a latent variable model and interacting with distributional word embeddings, to enrich the representation of tokens and measure the relatedness between tokens and each learnable cluster centroid. The proposed method jointly learns word clustering centroids and clustering-token alignments, achieving the state of the art results on multiple benchmark datasets and proving that the proposed cluster-token alignment mechanism is indeed favorable to text classification. Notably, our qualitative analysis has conspicuously illustrated that text representations learned by the proposed model are in accord well with our intuition. * Equal contribution. Preprint. Under review.

Findings of the Association for Computational Linguistics: EMNLP 2021, 2021

Generative Adversarial Networks (GANs) have achieved great success in image synthesis, but have p... more Generative Adversarial Networks (GANs) have achieved great success in image synthesis, but have proven to be difficult to generate natural language. Challenges arise from the uninformative learning signals passed from the discriminator. In other words, the poor learning signals limit the learning capacity for generating languages with rich structures and semantics. In this paper, we propose to adopt the counter-contrastive learning (CCL) method to support the generator's training in language GANs. In contrast to standard GANs that adopt a simple binary classifier to discriminate whether a sample is real or fake, we employ a counter-contrastive learning signal that advances the training of language synthesizers by (1) pulling the language representations of generated and real samples together and (2) pushing apart representations of real samples to compete with the discriminator and thus prevent the discriminator from being overtrained. We evaluate our method on both synthetic and real benchmarks and yield competitive performance compared to previous GANs for adversarial sequence generation.

Journal of Cleaner Production, 2021

Abstract Flotation has a great potential in the separation of complicated poly-mixtures to obtain... more Abstract Flotation has a great potential in the separation of complicated poly-mixtures to obtain qualified resins for recycling, but it remains a technological challenge for selective wetting between components. This study provided a combined flotation process with the prepositive plasma modification, separating hazardous polyvinyl chloride (PVC) from recyclable high-density polyethylene (HDPE) and polyethylene terephthalate (PET). The plasma process was characterized by surface morphology, molecular weight, contact angle, surface free energy and spectrum. The results suggested that amorphous low molecular weight oxidic substances can be selectively introduced on these polymers through the hydrogen extraction and Norrish І type radical pathways in plasma zone, leading to a desired surface wetting. Under the optimal unit activating energy of 12.0 kJ/m2 (PVC-HDPE) and 15.0 kJ/m2 (PVC-PET), separated PVC can reach above 93% of both recovery and purity. Benefiting from the green and mild operation in plasma unit, this study paves a new way to effectively separate waste plastics as an alternative to traditional surface modification with chemical reagents.

Proceedings of the Third Workshop on Multimodal Artificial Intelligence, 2021

Emotion recognition in conversation has received considerable attention recently because of its p... more Emotion recognition in conversation has received considerable attention recently because of its practical industrial applications. Existing methods tend to overlook the immediate mutual interaction between different speakers in the speaker-utterance level, or apply single speaker-agnostic RNN for utterances from different speakers. We propose COIN, a conversational interactive model to mitigate this problem by applying state mutual interaction within history contexts. In addition, we introduce a stacked global interaction module to capture the contextual and inter-dependency representation in a hierarchical manner. To improve the robustness and generalization during training, we generate adversarial examples by applying the minor perturbations on multimodal feature inputs, unveiling the benefits of adversarial examples for emotion detection. The proposed model empirically achieves the current state-of-the-art results on the IEMO-CAP benchmark dataset.