Search

In this study, we present the bicubic Hermite element method (BHEM), a new computational framework devised for the elastodynamic simulation of parametric thin-shell structures. The BHEM is constructed based on parametric quadrilateral Hermite patches, which serve as a unified representation for shell geometry, simulation, collision avoidance, as well as rendering. Compared with the commonly utilized linear FEM, the BHEM offers higher-order solution spaces, enabling the capture of more intricate and smoother geometries while employing significantly fewer finite elements. In comparison to other high-order methods, the BHEM achieves conforming $\mathcal{C}^1$ continuity for Kirchhoff-Love (KL) shells with minimal complexity. Furthermore, by leveraging the subdivision and convex hull properties of Hermite patches, we develop an efficient algorithm for ray-patch intersections, facilitating collision handling in simulations and ray tracing in rendering. This eliminates the need for laborious remodeling of the pre-existing parametric surface as the conventional approaches do. We substantiate our claims with comprehensive experiments, which demonstrate the high accuracy and versatility of the proposed method.

Despite CLIP being the foundation model in numerous vision-language applications, the CLIP suffers from a severe text spotting bias. Such bias causes CLIP models to `Parrot' the visual text embedded within images while disregarding the authentic visual semantics. We uncover that in the most popular image-text dataset LAION-2B, the captions also densely parrot (spell) the text embedded in images. Our analysis shows that around 50% of images are embedded with visual text content, and around 30% of captions words are in these embedded visual content. Based on such observation, we thoroughly inspect the different released versions of CLIP models and verify that the visual text is the dominant factor in measuring the LAION-style image-text similarity for these models. To examine whether these parrot captions shape the text spotting bias, we train a series of CLIP models with LAION subsets curated by different parrot-caption-oriented criteria. We show that training with parrot captions easily shapes such bias but harms the expected visual-language representation learning in CLIP models. This suggests that it is urgent to revisit either the design of CLIP-like models or the existing image-text dataset curation pipeline built on CLIP score filtering., Comment: project page: https://linyq17.github.io/CLIP-Parrot-Bias/. Add more analysis and ablation studies. Update Figure 3 with a more precise metric

In this paper, we formulate and analyse a geometric low-regularity integrator for solving the nonlinear Klein-Gordon equation in the $d$-dimensional space with $d=1,2,3$. The integrator is constructed based on the two-step trigonometric method and thus it has a simple form. Error estimates are rigorously presented to show that the integrator can achieve second-order time accuracy in the energy space under the regularity requirement in $H^{1+\frac{d}{4}}\times H^{\frac{d}{4}}$. Moreover, the time symmetry of the scheme ensures its good long-time energy conservation which is rigorously proved by the technique of modulated Fourier expansions. A numerical test is presented and the numerical results demonstrate the superiorities of the new integrator over some existing methods.

The numerical simulation of three-dimensional charged-particle dynamics (CPD) under strong magnetic field is a basic and challenging algorithmic task in plasma physics. In this paper, we introduce a new methodology to design two-scale exponential integrators for three-dimensional CPD whose magnetic field's strength is inversely proportional to a dimensionless and small parameter $0<\varepsilon \ll 1$. By dealing with the transformed form of three-dimensional CPD, we linearize the magnetic field and put the residual component in a new nonlinear function which is shown to be uniformly bounded. Based on this foundation and the proposed two-scale exponential integrators, a class of novel integrators is formulated and studied. The corresponding uniform accuracy of the proposed $r$-th order integrator is shown to be $\mathcal{O}(h^r)$, where $r=1,2,3,4$ and the constant symbolized by $\mathcal{O}$, the time stepsize $h$ and the computation cost are all independent of $\varepsilon$. Moreover, in the case of maximal ordering strong magnetic field, improved error bound $\mathcal{O}(\varepsilon^r h^r)$ is obtained for the proposed $r$-th order integrator. A rigorous proof of these uniform and improved error bounds is presented, and a numerical test is performed to illustrate the error and efficiency behaviour of the proposed integrators.

Hallucination, posed as a pervasive challenge of multi-modal large language models (MLLMs), has significantly impeded their real-world usage that demands precise judgment. Existing methods mitigate this issue with either training with specific designed data or inferencing with external knowledge from other sources, incurring inevitable additional costs. In this paper, we present OPERA, a novel MLLM decoding method grounded in an Over-trust Penalty and a Retrospection-Allocation strategy, serving as a nearly free lunch to alleviate the hallucination issue without additional data, knowledge, or training. Our approach begins with an interesting observation that, most hallucinations are closely tied to the knowledge aggregation patterns manifested in the self-attention matrix, i.e., MLLMs tend to generate new tokens by focusing on a few summary tokens, but not all the previous tokens. Such partial over-trust inclination results in the neglecting of image tokens and describes the image content with hallucination. Based on the observation, OPERA introduces a penalty term on the model logits during the beam-search decoding to mitigate the over-trust issue, along with a rollback strategy that retrospects the presence of summary tokens in the previously generated tokens, and re-allocate the token selection if necessary. With extensive experiments, OPERA shows significant hallucination-mitigating performance on different MLLMs and metrics, proving its effectiveness and generality. Our code is available at: https://github.com/shikiw/OPERA., Comment: CVPR 2024, code is available at https://github.com/shikiw/OPERA

Multimodal large language models have made significant advancements in recent years, yet they still suffer from a common issue known as the "hallucination problem", in which the models generate textual descriptions that inaccurately depict or entirely fabricate content from associated images. This paper introduces a novel solution, Hallucination-Aware Direct Preference Optimization (HA-DPO), which reframes the hallucination problem as a preference selection task. The model is trained to favor the non-hallucinating response when presented with two responses of the same image (one accurate and one hallucinatory). Furthermore, this paper proposes an efficient pipeline for constructing positive~(non-hallucinatory) and negative~(hallucinatory) sample pairs, ensuring a high-quality, style-consistent dataset for robust preference learning. When applied to three mainstream multimodal models, HA-DPO significantly reduced hallucination issues and amplified the models' generalization capabilities. Notably, the MiniGPT-4 model, when enhanced with HA-DPO, demonstrated a substantial improvement: POPE accuracy rose from 51.13% to 86.13% (an absolute improvement of 35%), and the MME score surged from 932.00 to 1326.46 (a relative improvement of 42.32%). The codes, models, and datasets are made accessible at https://opendatalab.github.io/HA-DPO., Comment: Project Website: https://opendatalab.github.io/HA-DPO, Code: https://github.com/opendatalab/HA-DPO

We propose a theoretical method to calculate the stellar $\beta$-decay rates of nuclei in stellar environments with high temperature and density, based on the projected shell model, where contributions from both allowed and first-forbidden transitions are taken into account. As the first example, the stellar $\beta$-decay rate of one of the last $s$-process branching-point nuclei, $^{204}$Tl, is calculated and studied, where all related transitions are first-forbidden transitions. For the terrestrial case, the ground-state to ground-state transition is unique first-forbidden transition, which is described reasonably by our calculations. At the typical $s$-process temperature ($T\approx 0.3$ GK), non-unique first-forbidden transitions from thermally populated excited states of the parent nucleus are involved, the effective rate from our calculations is much lower than the one from the widely used data tables by Takahashi and Yokoi. Effect of the quenching factors for nuclear matrix elements in first-forbidden transitions on the stellar $\beta$-decay rates is discussed as well., Comment: With 9 pages, 6 figures. Comments and criticisms are welcome

Simultaneous machine translation (SiMT) is a challenging task that requires starting translation before the full source sentence is available. Prefix-to-prefix framework is often applied to SiMT, which learns to predict target tokens using only a partial source prefix. However, due to the word order difference between languages, misaligned prefix pairs would make SiMT models suffer from serious hallucination problems, i.e. target outputs that are unfaithful to source inputs. Such problems can not only produce target tokens that are not supported by the source prefix, but also hinder generating the correct translation by receiving more source words. In this work, we propose a Confidence-Based Simultaneous Machine Translation (CBSiMT) framework, which uses model confidence to perceive hallucination tokens and mitigates their negative impact with weighted prefix-to-prefix training. Specifically, token-level and sentence-level weights are calculated based on model confidence and acted on the loss function. We explicitly quantify the faithfulness of the generated target tokens using the token-level weight, and employ the sentence-level weight to alleviate the disturbance of sentence pairs with serious word order differences on the model. Experimental results on MuST-C English-to-Chinese and WMT15 German-to-English SiMT tasks demonstrate that our method can consistently improve translation quality at most latency regimes, with up to 2 BLEU scores improvement at low latency.

The first-forbidden transition of nuclear $\beta$ decay is expected to play crucial roles in many aspects in nuclear physics, nuclear astrophysics and particle physics such as the stellar $\beta$-decay rates and the reactor anti-neutrino spectra. In this work we develop the projected shell model (PSM) for description of first-forbidden transition of nuclear $\beta$ decay for the first time. Detailed theoretical framework and logics are provided, and 35 dominant first-forbidden transitions that are expected to be important for the reactor anti-neutrino spectra problems are calculated and compared systematically with the data to test the new development of the PSM. The corresponding experimental Log$f_0 t$ values are described reasonably, and the quenching factors of nuclear matrix elements are found to affect the Log$f_0 t$ values as well as the related shape factors, which may be helpful for better understanding of the reactor anti-neutrino spectra problems.

Recent advances in large language models (LLMs) have revolutionized the landscape of reasoning tasks. To enhance the capabilities of LLMs to emulate human reasoning, prior studies have focused on modeling reasoning steps using various thought structures like chains, trees, or graphs. However, LLM-based reasoning still encounters the following challenges: (1) Limited adaptability of preset structures to diverse tasks; (2) Insufficient precision in exploiting known conditions to derive new ones; and (3) Inadequate consideration of historical reasoning experiences for subsequent reasoning steps. To this end, we propose DetermLR, a novel perspective that rethinks the reasoning process as an evolution from indeterminacy to determinacy. First, we categorize known conditions into two types: determinate and indeterminate premises This provides an oveall direction for the reasoning process and guides LLMs in converting indeterminate data into progressively determinate insights. Subsequently, we leverage quantitative measurements to prioritize more relevant premises to explore new insights. Furthermore, we automate the storage and extraction of available premises and reasoning paths with reasoning memory, preserving historical reasoning details for subsequent reasoning steps. Comprehensive experimental results demonstrate that DetermLR surpasses all baselines on various logical reasoning benchmarks: LogiQA, ProofWriter, FOLIO, PrOntoQA, and LogicalDeduction. Compared to previous multi-step reasoning methods, DetermLR achieves higher accuracy with fewer reasoning steps, highlighting its superior efficiency and effectiveness in solving logical reasoning tasks., Comment: Accepted at ACL 2024 Main, Code repo: https://github.com/XiaoMi/DetermLR

Blind super-resolution (BSR) methods based on high-resolution (HR) reconstruction codebooks have achieved promising results in recent years. However, we find that a codebook based on HR reconstruction may not effectively capture the complex correlations between low-resolution (LR) and HR images. In detail, multiple HR images may produce similar LR versions due to complex blind degradations, causing the HR-dependent only codebooks having limited texture diversity when faced with confusing LR inputs. To alleviate this problem, we propose the Rich Texture-aware Codebook-based Network (RTCNet), which consists of the Degradation-robust Texture Prior Module (DTPM) and the Patch-aware Texture Prior Module (PTPM). DTPM effectively mines the cross-resolution correlation of textures between LR and HR images by exploiting the cross-resolution correspondence of textures. PTPM uses patch-wise semantic pre-training to correct the misperception of texture similarity in the high-level semantic regularization. By taking advantage of this, RTCNet effectively gets rid of the misalignment of confusing textures between HR and LR in the BSR scenarios. Experiments show that RTCNet outperforms state-of-the-art methods on various benchmarks by up to 0.16 ~ 0.46dB.

Deep neural networks (DNNs) are vulnerable to adversarial examples crafted by well-designed perturbations. This could lead to disastrous results on critical applications such as self-driving cars, surveillance security, and medical diagnosis. At present, adversarial training is one of the most effective defenses against adversarial examples. However, traditional adversarial training makes it difficult to achieve a good trade-off between clean accuracy and robustness since spurious features are still learned by DNNs. The intrinsic reason is that traditional adversarial training makes it difficult to fully learn core features from adversarial examples when adversarial noise and clean examples cannot be disentangled. In this paper, we disentangle the adversarial examples into natural and perturbed patterns by bit-plane slicing. We assume the higher bit-planes represent natural patterns and the lower bit-planes represent perturbed patterns, respectively. We propose a Feature-Focusing Adversarial Training (F$^2$AT), which differs from previous work in that it enforces the model to focus on the core features from natural patterns and reduce the impact of spurious features from perturbed patterns. The experimental results demonstrated that F$^2$AT outperforms state-of-the-art methods in clean accuracy and adversarial robustness.

Large-scale, big-variant, and high-quality data are crucial for developing robust and successful deep-learning models for medical applications since they potentially enable better generalization performance and avoid overfitting. However, the scarcity of high-quality labeled data always presents significant challenges. This paper proposes a novel approach to address this challenge by developing controllable diffusion models for medical image synthesis, called EMIT-Diff. We leverage recent diffusion probabilistic models to generate realistic and diverse synthetic medical image data that preserve the essential characteristics of the original medical images by incorporating edge information of objects to guide the synthesis process. In our approach, we ensure that the synthesized samples adhere to medically relevant constraints and preserve the underlying structure of imaging data. Due to the random sampling process by the diffusion model, we can generate an arbitrary number of synthetic images with diverse appearances. To validate the effectiveness of our proposed method, we conduct an extensive set of medical image segmentation experiments on multiple datasets, including Ultrasound breast (+13.87%), CT spleen (+0.38%), and MRI prostate (+7.78%), achieving significant improvements over the baseline segmentation methods. For the first time, to our best knowledge, the promising results demonstrate the effectiveness of our EMIT-Diff for medical image segmentation tasks and show the feasibility of introducing a first-ever text-guided diffusion model for general medical image segmentation tasks. With carefully designed ablation experiments, we investigate the influence of various data augmentation ratios, hyper-parameter settings, patch size for generating random merging mask settings, and combined influence with different network architectures., Comment: 15 pages

Conventional dialogue summarization methods directly generate summaries and do not consider user's specific interests. This poses challenges in cases where the users are more focused on particular topics or aspects. With the advancement of instruction-finetuned language models, we introduce instruction-tuning to dialogues to expand the capability set of dialogue summarization models. To overcome the scarcity of instructive dialogue summarization data, we propose a three-step approach to synthesize high-quality query-based summarization triples. This process involves summary-anchored query generation, query filtering, and query-based summary generation. By training a unified model called InstructDS (Instructive Dialogue Summarization) on three summarization datasets with multi-purpose instructive triples, we expand the capability of dialogue summarization models. We evaluate our method on four datasets, including dialogue summarization and dialogue reading comprehension. Experimental results show that our approach outperforms the state-of-the-art models and even models with larger sizes. Additionally, our model exhibits higher generalizability and faithfulness, as confirmed by human subjective evaluations., Comment: EMNLP 2023 Main Conference - Summarization (update for acknowledgement)