Towards Boosting the Open-Domain Chatbot with Human Feedback

Many open-domain dialogue models pre-trained with social media comments can generate coherent replies but have difficulties producing engaging responses when interacting with real users. This phenomenon might mainly result from the deficiency of annotated human-human conversations and the misalignment with human preference. In this paper, we propose a novel and efficient approach Diamante to boost the open-domain chatbot, where two kinds of human feedback (including explicit demonstration and implicit preference) are collected and leveraged. By asking annotators to select or amend the model-generated candidate responses, Diamante efficiently collects the human demonstrated responses and constructs a Chinese chit-chat dataset. To enhance the alignment with human preference, Diamante leverages the implicit preference in the data collection process and introduces the generation-evaluation joint training. Comprehensive experiments indicate that the Diamante dataset and joint training paradigm can significantly boost the performance of Chinese pre-trained dialogue models.

Semi-supervised Semantic Segmentation with Mutual Knowledge Distillation

Consistency regularization has been widely studied in recent semi-supervised semantic segmentation methods. Remarkable performance has been achieved, benefiting from image, feature, and network perturbations. To make full use of these perturbations, in this work, we propose a new consistency regularization framework called mutual knowledge distillation (MKD). We innovatively introduce two auxiliary mean-teacher models based on the consistency regularization method. More specifically, we use the pseudo label generated by one mean teacher to supervise the other student network to achieve a mutual knowledge distillation between two branches. In addition to using image-level strong and weak augmentation, we also employ feature augmentation considering implicit semantic distributions to add further perturbations to the students. The proposed framework significantly increases the diversity of the training samples. Extensive experiments on public benchmarks show that our framework outperforms previous state-of-the-art(SOTA) methods under various semi-supervised settings. Code is available at: https://github.com/jianlong-yuan/semi-mmseg.

FAKD: Feature Augmented Knowledge Distillation for Semantic Segmentation

In this work, we explore data augmentations for knowledge distillation on semantic segmentation. To avoid over-fitting to the noise in the teacher network, a large number of training examples is essential for knowledge distillation. Imagelevel argumentation techniques like flipping, translation or rotation are widely used in previous knowledge distillation framework. Inspired by the recent progress on semantic directions on feature-space, we propose to include augmentations in feature space for efficient distillation. Specifically, given a semantic direction, an infinite number of augmentations can be obtained for the student in the feature space. Furthermore, the analysis shows that those augmentations can be optimized simultaneously by minimizing an upper bound for the losses defined by augmentations. Based on the observation, a new algorithm is developed for knowledge distillation in semantic segmentation. Extensive experiments on four semantic segmentation benchmarks demonstrate that the proposed method can boost the performance of current knowledge distillation methods without any significant overhead. Code is available at: https://github.com/jianlong-yuan/FAKD.

GEM-2: Next Generation Molecular Property Prediction Network with Many-body and Full-range Interaction Modeling

Molecular property prediction is a fundamental task in the drug and material industries. Physically, the properties of a molecule are determined by its own electronic structure, which can be exactly described by the Schr\"odinger equation. However, solving the Schr\"odinger equation for most molecules is extremely challenging due to long-range interactions in the behavior of a quantum many-body system. While deep learning methods have proven to be effective in molecular property prediction, we design a novel method, namely GEM-2, which comprehensively considers both the long-range and many-body interactions in molecules. GEM-2 consists of two interacted tracks: an atom-level track modeling both the local and global correlation between any two atoms, and a pair-level track modeling the correlation between all atom pairs, which embed information between any 3 or 4 atoms. Extensive experiments demonstrated the superiority of GEM-2 over multiple baseline methods in quantum chemistry and drug discovery tasks.

HelixFold-Single: MSA-free Protein Structure Prediction by Using Protein Language Model as an Alternative

AI-based protein structure prediction pipelines, such as AlphaFold2, have achieved near-experimental accuracy. These advanced pipelines mainly rely on Multiple Sequence Alignments (MSAs) as inputs to learn the co-evolution information from the homologous sequences. Nonetheless, searching MSAs from protein databases is time-consuming, usually taking dozens of minutes. Consequently, we attempt to explore the limits of fast protein structure prediction by using only primary sequences of proteins. HelixFold-Single is proposed to combine a large-scale protein language model with the superior geometric learning capability of AlphaFold2. Our proposed method, HelixFold-Single, first pre-trains a large-scale protein language model (PLM) with thousands of millions of primary sequences utilizing the self-supervised learning paradigm, which will be used as an alternative to MSAs for learning the co-evolution information. Then, by combining the pre-trained PLM and the essential components of AlphaFold2, we obtain an end-to-end differentiable model to predict the 3D coordinates of atoms from only the primary sequence. HelixFold-Single is validated in datasets CASP14 and CAMEO, achieving competitive accuracy with the MSA-based methods on the targets with large homologous families. Furthermore, HelixFold-Single consumes much less time than the mainstream pipelines for protein structure prediction, demonstrating its potential in tasks requiring many predictions. The code of HelixFold-Single is available at https://github.com/PaddlePaddle/PaddleHelix/tree/dev/apps/protein_folding/helixfold-single, and we also provide stable web services on https://paddlehelix.baidu.com/app/drug/protein-single/forecast.

HelixFold: An Efficient Implementation of AlphaFold2 using PaddlePaddle

Accurate protein structure prediction can significantly accelerate the development of life science. The accuracy of AlphaFold2, a frontier end-to-end structure prediction system, is already close to that of the experimental determination techniques. Due to the complex model architecture and large memory consumption, it requires lots of computational resources and time to implement the training and inference of AlphaFold2 from scratch. The cost of running the original AlphaFold2 is expensive for most individuals and institutions. Therefore, reducing this cost could accelerate the development of life science. We implement AlphaFold2 using PaddlePaddle, namely HelixFold, to improve training and inference speed and reduce memory consumption. The performance is improved by operator fusion, tensor fusion, and hybrid parallelism computation, while the memory is optimized through Recompute, BFloat16, and memory read/write in-place. Compared with the original AlphaFold2 (implemented with Jax) and OpenFold (implemented with PyTorch), HelixFold needs only 7.5 days to complete the full end-to-end training and only 5.3 days when using hybrid parallelism, while both AlphaFold2 and OpenFold take about 11 days. HelixFold saves 1x training time. We verified that HelixFold's accuracy could be on par with AlphaFold2 on the CASP14 and CAMEO datasets. HelixFold's code is available on GitHub for free download: https://github.com/PaddlePaddle/PaddleHelix/tree/dev/apps/protein_folding/helixfold, and we also provide stable web services on https://paddlehelix.baidu.com/app/drug/protein/forecast.

Dynamic Gradient Reactivation for Backward Compatible Person Re-identification

We study the backward compatible problem for person re-identification (Re-ID), which aims to constrain the features of an updated new model to be comparable with the existing features from the old model in galleries. Most of the existing works adopt distillation-based methods, which focus on pushing new features to imitate the distribution of the old ones. However, the distillation-based methods are intrinsically sub-optimal since it forces the new feature space to imitate the inferior old feature space. To address this issue, we propose the Ranking-based Backward Compatible Learning (RBCL), which directly optimizes the ranking metric between new features and old features. Different from previous methods, RBCL only pushes the new features to find best-ranking positions in the old feature space instead of strictly alignment, and is in line with the ultimate goal of backward retrieval. However, the sharp sigmoid function used to make the ranking metric differentiable also incurs the gradient vanish issue, therefore stems the ranking refinement during the later period of training. To address this issue, we propose the Dynamic Gradient Reactivation (DGR), which can reactivate the suppressed gradients by adding dynamic computed constant during forward step. To further help targeting the best-ranking positions, we include the Neighbor Context Agents (NCAs) to approximate the entire old feature space during training. Unlike previous works which only test on the in-domain settings, we make the first attempt to introduce the cross-domain settings (including both supervised and unsupervised), which are more meaningful and difficult. The experimental results on all five settings show that the proposed RBCL outperforms previous state-of-the-art methods by large margins under all settings.

TCR: A Transformer Based Deep Network for Predicting Cancer Drugs Response

Predicting clinical outcomes to anti-cancer drugs on a personalized basis is challenging in cancer treatment due to the heterogeneity of tumors. Traditional computational efforts have been made to model the effect of drug response on individual samples depicted by their molecular profile, yet overfitting occurs because of the high dimension for omics data, hindering models from clinical application. Recent research shows that deep learning is a promising approach to build drug response models by learning alignment patterns between drugs and samples. However, existing studies employed the simple feature fusion strategy and only considered the drug features as a whole representation while ignoring the substructure information that may play a vital role when aligning drugs and genes. Hereby in this paper, we propose TCR (Transformer based network for Cancer drug Response) to predict anti-cancer drug response. By utilizing an attention mechanism, TCR is able to learn the interactions between drug atom/sub-structure and molecular signatures efficiently in our study. Furthermore, a dual loss function and cross sampling strategy were designed to improve the prediction power of TCR. We show that TCR outperformed all other methods under various data splitting strategies on all evaluation matrices (some with significant improvement). Extensive experiments demonstrate that TCR shows significantly improved generalization ability on independent in-vitro experiments and in-vivo real patient data. Our study highlights the prediction power of TCR and its potential value for cancer drug repurpose and precision oncology treatment.