{"arxiv_id":"1706.03762","incoming":{"papers":4153,"sentences":6845,"groups":[{"kind":"contests","label":"Contests it","papers":1,"sentences":1,"quotes":[{"arxiv_id":"2105.14039","title":"Towards mental time travel: a hierarchical memory for reinforcement learning agents","url":"/paper/2105.14039","section":"Introduction","sentence":"While LSTMs [ref] , Transformers [ref] , and variants like the TransformerXL [ref] can serve as RL agent memories in short tasks [ref] , we show that they are ineffective at detailed recall even after a few minutes.","intent_p":0.39,"contests_p":0.78}]},{"kind":"extends","label":"Builds on it","papers":17,"sentences":17,"quotes":[{"arxiv_id":"2109.05388","title":"The Impact of Positional Encodings on Multilingual Compression","url":"/paper/2109.05388","section":"Introduction","sentence":"Contributions We build on [ref] and demonstrate, through a series of experiments on synthetic and real data, that the choice of positional encoding mechanism has a significant effect on cross-lingual model performance: While many positional encodings have been proposed in monolingual settings as improvements over sinusoidal or absolute positional encodings, originally proposed in [ref] and [ref] , including untied positional encodings (TUPE; [ref] ) and relative positional encodings [ref] , none of these better facilitate cross-lingual compression or sharing.","intent_p":0.82,"contests_p":null},{"arxiv_id":"2303.16097","title":"Attention Boosted Autoencoder for Building Energy Anomaly Detection","url":"/paper/2303.16097","section":"TiFe Attention Model","sentence":"Extending from the scaled dot product attention introduced in [ref] , we define an attention (referred to as TiFe Attention) mechanism that captures relationships across time and feature dimensions as shown in Figure[ref].","intent_p":0.99,"contests_p":null},{"arxiv_id":"2206.04003","title":"Patch-based Object-centric Transformers for Efficient Video Generation","url":"/paper/2206.04003","section":"Introduction","sentence":"We build upon prior work in video generation architectures using autoregressive discrete latent variable models that combine VQ-VAE [ref] and transformers [ref] by introducing a novel two-stream attention module that jointly models VQ-VAE discrete image latents along with predicting object bounding boxes for each frame.","intent_p":0.95,"contests_p":null},{"arxiv_id":"2311.10642","title":"Rethinking Attention: Exploring Shallow Feed-Forward Neural Networks as an Alternative to Attention Layers in Transformers","url":"/paper/2311.10642","section":"Introduction","sentence":"Inspired by prior work [ref] , [ref] which explore the feasibility of training shallow feed-forward networks to emulate the behavior of deep convolutional networks with deep networks as teachers, we conduct a similar investigation on the original Transformer presented in [ref] .","intent_p":0.93,"contests_p":null},{"arxiv_id":"1710.07654","title":"Deep Voice 3: Scaling Text-to-Speech with Convolutional Sequence Learning","url":"/paper/1710.07654","section":"Related Work","sentence":"Deep Voice 3 also builds upon the convolutional sequence-to-sequence architecture from [ref] by introducing a positional encoding similar to that used in [ref] , augmented with a rate adjustment to account for the mismatch between input and output domain lengths.","intent_p":0.99,"contests_p":null}]},{"kind":"method_or_tool","label":"Uses its method or tool","papers":1173,"sentences":1542,"quotes":[{"arxiv_id":"1905.06638","title":"Latent Universal Task-Specific BERT","url":"/paper/1905.06638","section":"Universal Modification","sentence":"In order to improve performance and reduce the number of model parameters, we have replaced the Transformer [ref] model with a design, similar to the Universal Transformer [ref] , where the sequential self-attention blocks of the Transformer Encoder are replaced with a single recurring block.","intent_p":0.83,"contests_p":null},{"arxiv_id":"2211.01736","title":"Transformers on Multilingual Clause-Level Morphology","url":"/paper/2211.01736","section":"Conclusion","sentence":"For inflection we used a vanilla Transformer adapted from [ref] and applying data hallucination substantially improves accuracy [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2509.11233","title":"TransZero: Parallel Tree Expansion in MuZero using Transformer Networks","url":"/paper/2509.11233","section":"Introduction","sentence":"Our approach has two main components: (i) a transformer-based [ref] dynamics network that generates entire rollouts in parallel via self-attention, and (ii) a Mean-Variance Constrained (MVC) evaluator [ref] that enables node expansion independent of visitation counts.","intent_p":0.95,"contests_p":null},{"arxiv_id":"2107.05944","title":"The Piano Inpainting Application","url":"/paper/2107.05944","section":"Efficient masking scheme","sentence":"In this work, we use Linear Transformers [ref] as they provide a way to overcome the $O(T^2)$ complexity of computing the self-attention, which allows us to consider longer sequences, and we rely on a standard encoder-decoder architecture with transformers as proposed in [ref] .","intent_p":0.99,"contests_p":null},{"arxiv_id":"2010.07773","title":"NUIG-Shubhanker@Dravidian-CodeMix-FIRE2020: Sentiment Analysis of Code-Mixed Dravidian text using XLNet","url":"/paper/2010.07773","section":"Introduction","sentence":"In this paper, we address these challenges using XLNet [ref] [XLNet is an auto-regressive model [ref] which is built with the transformer architecture [ref] with a two stream attention mechanism [ref] .","intent_p":0.97,"contests_p":null}]},{"kind":"data_used","label":"Uses its data or values","papers":9,"sentences":11,"quotes":[{"arxiv_id":"2309.16716","title":"Towards Safe Autonomy in Hybrid Traffic: Detecting Unpredictable Abnormal Behaviors of Human Drivers via Information Sharing","url":"/paper/2309.16716","section":"Introduction","sentence":"We evaluate our proposed MEATP on NGSIM [ref] and Argoverse [ref] datasets, the experiment results show that MEATP with information sharing outperforms the well-adopted Long-Short Term Memory (LSTM) [ref] and some transformer [ref] based trajectory predictors.","intent_p":0.8,"contests_p":null},{"arxiv_id":"2308.12157","title":"Evaluation of Faithfulness Using the Longest Supported Subsequence","url":"/paper/2308.12157","section":"Summarization","sentence":"We employ the 2,000 annotations released by [ref] , which contain human ratings of factuality for summaries generated by four neural models: pointer generator network (“PG”: [ref] ), Topic-aware convolutional sequence-to-sequence model (“TCONV”: [ref] ), a transformer-based model (“TRANS2S”: [ref] ), and BERT ( [ref] , [ref] .","intent_p":0.95,"contests_p":null},{"arxiv_id":"2505.04652","title":"Rethinking Boundary Detection in Deep Learning-Based Medical Image Segmentation","url":"/paper/2505.04652","section":"Datasets and Evaluation Metrics","sentence":"In our experiments, we evaluate the performance of CTO on seven publicly available MedISeg datasets, namely ISIC 2016 [ref] , ISIC 2018 [ref] , PH2 [ref] , the Colon Nuclei Identification and Counting (CoNIC) challenge dataset [ref] , the Liver Tumor Segmentation (LiTS17) challenge dataset [ref] , MSD BraTS [ref] , and the Beyond the Cranial Vault (BTCV) challenge dataset [ref] .","intent_p":0.98,"contests_p":null},{"arxiv_id":"1810.10126","title":"Area Attention","url":"/paper/1810.10126","section":"Token-Level Neural Machine Translation","sentence":"We use the same dataset as the one used in [ref] in which the WMT 2014 English-German (EN-DE) dataset contains about 4.5 million English-German sentence pairs, and the English-French (EN-FR) dataset has about 36 million English-French sentence pairs [ref] .","intent_p":0.98,"contests_p":null},{"arxiv_id":"2607.04535","title":"ManifoldFlow: Spectrum-Learnable Stiefel Optimization via SPD Relaxation","url":"/paper/2607.04535","section":"Experiments","sentence":"We therefore present the results by task family rather than as a flat list of endpoints: recurrent language models on WikiText, tabular MLPs on UCI benchmarks, image-derived MLPs on Fashion-MNIST and CIFAR, a Mini-Transformer feed-forward block, and convolutional classifier heads on CIFAR [ref] .","intent_p":0.86,"contests_p":null}]},{"kind":"comparison","label":"Compares with it","papers":58,"sentences":64,"quotes":[{"arxiv_id":"2406.14189","title":"In Tree Structure Should Sentence Be Generated","url":"/paper/2406.14189","section":"Experiments and Evaluation","sentence":"Results are reported in <Ref>, and as we can see, the proposed approach gives better BLEU scores compared to its baseline alternative [ref] on the WMT 2014 English-to-German translation task [ref] .","intent_p":0.8,"contests_p":null},{"arxiv_id":"2503.10492","title":"Meta-learning characteristics and dynamics of quantum systems","url":"/paper/2503.10492","section":"Introduction","sentence":"MALGO outperforms state-of-the-art meta-learning techniques [ref] , a vanilla transformer [ref] , and a multilayer perceptron (MLP) in predicting the dynamics of three different families of quantum systems: a closed and open two-level system (TLS), and a many-body Heisenberg system.","intent_p":0.87,"contests_p":null},{"arxiv_id":"2205.06457","title":"ViT5: Pretrained Text-to-Text Transformer for Vietnamese Language Generation","url":"/paper/2205.06457","section":"Baselines","sentence":"In order to verify the effectiveness of our proposed methods, we compare ViT5 models with the Transformer models based on [ref] , the ViSum BERT2BERT models [ref] , multilingual encoder-decoder model [ref] , and Vietnamese encoder-decoder BARTpho model [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"2108.04962","title":"Adaptive Multi-Resolution Attention with Linear Complexity","url":"/paper/2108.04962","section":"Experiment Settings","sentence":"Aside from the vanilla Transformer [ref] , we compare our model against other efficient self-attention variants, including Reformer [ref] , Linear Transformer [ref] , Performer [ref] , Linformer [ref] , Big Bird [ref] and Nyströmformer [ref] .","intent_p":0.98,"contests_p":null},{"arxiv_id":"2508.15010","title":"TOAST: Fast and scalable auto-partitioning based on principled static analysis","url":"/paper/2508.15010","section":"Introduction","sentence":"We demonstrate that for practically relevant large ML models ( Transformer [ref] , Graph Network Simulator (GNS) [ref] , U-Net [ref] ) discovers partitionings that are (a) better than best known manual partitioning strategies and (b) at least as good, and usually better, than partitionings discovered by the state-of-the-art Alpa auto-partitioner [ref] .","intent_p":0.81,"contests_p":null}]},{"kind":"supports_this_paper","label":"Cites it as support","papers":21,"sentences":21,"quotes":[{"arxiv_id":"2107.05264","title":"The Brownian motion in the transformer model","url":"/paper/2107.05264","section":"The Brownian motion in MHSA module","sentence":"As pointed in [ref] , each components of token $v_i$ is independent random variable, so $v_i$ is a random variables in $d$ dimensional feature space. * After layer normalization, these variables are points on the hyper-sphere.","intent_p":0.99,"contests_p":null},{"arxiv_id":"2209.08055","title":"A Transformer-Based Approach for Improving App Review Response Generation","url":"/paper/2209.08055","section":"Main Results","sentence":"As demonstrated in previous studies in natural language processing [ref] , the self-attention mechanism in the Transformers can help the model capture the long-term dependence information better.","intent_p":0.99,"contests_p":null},{"arxiv_id":"1910.11162","title":"U-Time: A Fully Convolutional Network for Time Series Segmentation Applied to Sleep Staging","url":"/paper/1910.11162","section":"Introduction","sentence":"While recurrent neural networks are conceptually appealing for time series analysis, they are often difficult to tune and optimize in practice, and it has been found that for many tasks across domains recurrent models can be replaced by feed-forward systems without sacrificing accuracy [ref] .","intent_p":0.98,"contests_p":null},{"arxiv_id":"2512.17375","title":"AdvJudge-Zero: Binary Decision Flips in LLM-as-a-Judge via Adversarial Control Tokens","url":"/paper/2512.17375","section":"Conclusion","sentence":"Because the transformer [ref] body appears empirically near-injective in line with recent work [ref] , small prompt edits can move the final hidden state across this gate without erasing the underlying task semantics.","intent_p":0.98,"contests_p":null},{"arxiv_id":"2605.24949","title":"APT-Agent: Automated Penetration Testing using Large Language Models","url":"/paper/2605.24949","section":"Introduction","sentence":"This shortfall matches evidence that attention-based models emphasize recent tokens and struggle with long-range procedural state [ref] .","intent_p":0.99,"contests_p":null}]},{"kind":"background","label":"Cites it as background","papers":2143,"sentences":2649,"quotes":[{"arxiv_id":"1905.06638","title":"Latent Universal Task-Specific BERT","url":"/paper/1905.06638","section":"Introduction","sentence":"Amongst the most notable of these models is BERT [ref] , which relies on the Transformer Encoder architecture [ref] and a unique pre-training objective known as “masked language model” (MLM).","intent_p":0.93,"contests_p":null},{"arxiv_id":"1910.13162","title":"An Efficient Model for Sentiment Analysis of Electronic Product Reviews in Vietnamese","url":"/paper/1910.13162","section":"Background","sentence":"In 2017, Vaswani et al. first introduced Self-attention Neural Network [ref] .","intent_p":0.96,"contests_p":null},{"arxiv_id":"2311.10642","title":"Rethinking Attention: Exploring Shallow Feed-Forward Neural Networks as an Alternative to Attention Layers in Transformers","url":"/paper/2311.10642","section":"Introduction","sentence":"The seminal paper [ref] which introduced the Transformer model has fundamentally altered the landscape of sequence-to-sequence modeling tasks.","intent_p":1.0,"contests_p":null},{"arxiv_id":"1711.02132","title":"Weighted Transformer Network for Machine Translation","url":"/paper/1711.02132","section":"Introduction","sentence":"In [ref] , the authors introduce the Transformer network, a novel architecture that avoids the recurrence equation and maps the input sequences into hidden states solely using attention.","intent_p":0.91,"contests_p":null},{"arxiv_id":"2106.00557","title":"Towards Interpretable Attention Networks for Cervical Cancer Analysis","url":"/paper/2106.00557","section":"Residual Attention Networks","sentence":"The residual attention network [ref] is a convolutional neural network based on the attention mechanism of [ref] .","intent_p":0.9,"contests_p":null}]}],"unlabelled_sentences":2540},"outgoing":{"papers":15,"sentences":27,"groups":[{"kind":"method_or_tool","label":"Uses the method or tool of","papers":3,"sentences":4,"quotes":[{"arxiv_id":"1609.08144","title":"Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation","url":"/paper/1609.08144","section":"Machine Translation","sentence":"We used beam search with a beam size of $4$ and length penalty $α=0.6$ [ref] .","intent_p":0.98,"contests_p":null},{"arxiv_id":"1607.06450","title":"Layer Normalization","url":"/paper/1607.06450","section":"Encoder and Decoder Stacks","sentence":"We employ a residual connection [ref] around each of the two sub-layers, followed by layer normalization [ref] .","intent_p":1.0,"contests_p":null},{"arxiv_id":"1608.05859","title":"Using the Output Embedding to Improve Language Models","url":"/paper/1608.05859","section":"Embeddings and Softmax","sentence":"In our model, we share the same weight matrix between the two embedding layers and the pre-softmax linear transformation, similar to [ref] .","intent_p":0.96,"contests_p":null}]},{"kind":"supports_this_paper","label":"Cites as support","papers":1,"sentences":1,"quotes":[{"arxiv_id":"1610.10099","title":"Neural Machine Translation in Linear Time","url":"/paper/1610.10099","section":"Why Self-Attention","sentence":"Doing so requires a stack of $O(n/k)$ convolutional layers in the case of contiguous kernels, or $O(log_k(n))$ in the case of dilated convolutions [ref] , increasing the length of the longest paths between any two positions in the network.","intent_p":0.97,"contests_p":null}]},{"kind":"background","label":"Cites as background","papers":4,"sentences":4,"quotes":[{"arxiv_id":"1602.02410","title":"Exploring the Limits of Language Modeling","url":"/paper/1602.02410","section":"Introduction","sentence":"Numerous efforts have since continued to push the boundaries of recurrent language models and encoder-decoder architectures [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"1609.08144","title":"Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation","url":"/paper/1609.08144","section":"Introduction","sentence":"Numerous efforts have since continued to push the boundaries of recurrent language models and encoder-decoder architectures [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"1508.04025","title":"Effective Approaches to Attention-based Neural Machine Translation","url":"/paper/1508.04025","section":"Introduction","sentence":"Numerous efforts have since continued to push the boundaries of recurrent language models and encoder-decoder architectures [ref] .","intent_p":0.97,"contests_p":null},{"arxiv_id":"1705.03122","title":"Convolutional Sequence to Sequence Learning","url":"/paper/1705.03122","section":"Positional Encoding","sentence":"There are many choices of positional encodings, learned and fixed [ref] .","intent_p":0.92,"contests_p":null}]}],"unlabelled_sentences":18},"claims":[],"thresholds":{"intent_p":0.8,"supports_p":0.95,"contests_p":0.7},"method":"Citing sentences come from each paper's LaTeX source; every \\cite marks the sentence and the works it cites. Pith's classifier typed each sentence. A label is shown at intent probability 0.8 or more (0.95 for support), a contest at 0.7 or more. Measured accuracy against two blind model families is at https://pith.science/accuracy. The quoted sentences are the citing authors' own.","held_up":{"arxiv_id":"1706.03762","kind":"model","paper_version":7,"prompt_version":"held-up-v3.5","generated_at":"2026-10-27T09:19:58+00:00","loaded_at":"2026-10-28T01:01:39.587845+00:00","evidence":"substantial","account":[],"groups":[{"kind":"method_or_tool","label":"Uses its method or tool","papers":13,"entries":[{"p":1.0,"quote":{"text":"In this work, we propose a novel approach for detecting energy theft in residential solar generation by leveraging multi-scale Convolutional Neural Network (CNN) [ref] , Long Short-Term Memory (LSTM) [ref] , and Transformer [ref] .","citing":"2505.18755","section":"Methodology"},"citing":[{"self":false,"year":2025,"title":"Smart Energy Guardian: A Hybrid Deep Learning Model for Detecting Fraudulent PV Generation","arxiv_id":"2505.18755"}],"verified":true,"paraphrase":"Used a Transformer in its approach for detecting energy theft in residential solar generation."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2018,"title":"A Self-Attention Network for Hierarchical Data Structures with an Application to Claims Management","arxiv_id":"1808.10543"}],"verified":true,"paraphrase":"Followed this paper by applying scaled dot product self-attention to turn a sequence of inputs into a sequence of features."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2020,"title":"Correcting the Autocorrect: Context-Aware Typographical Error Correction via Training Data Augmentation","arxiv_id":"2005.01158"}],"verified":true,"paraphrase":"Relied on a Transformer model to transform a noisy sequence with errors to the correct sequence."},{"p":1.0,"quote":{"text":"We employ a hybrid backbone with four Mamba layers [ref] and one Transformer layer [ref] .","citing":"2602.10603","section":"Architecture"},"citing":[{"self":false,"year":2026,"title":"dnaHNet: A Scalable and Hierarchical Foundation Model for Genomic Sequence Learning","arxiv_id":"2602.10603"}],"verified":true,"paraphrase":"Employed a hybrid backbone with four Mamba layers and one Transformer layer."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2017,"title":"Attending to All Mention Pairs for Full Abstract Biological Relation Extraction","arxiv_id":"1710.08312"}],"verified":true,"paraphrase":"Used the Transformer self-attention model to encode tokens by aggregating over their context in the entire sequence."},{"p":1.0,"quote":{"text":"We use Transformer encodings [ref] to capture the nuances of the language and uses neural attention mechanisms [ref] to learn to identify key sentences and words in the citation context of a citation.","citing":"2201.08089","section":"Multi-Module Attention Based Baseline Classifier"},"citing":[{"self":false,"year":2022,"title":"Why Did You Not Compare With That? Identifying Papers for Use as Baselines","arxiv_id":"2201.08089"}],"verified":true,"paraphrase":"Used Transformer encodings to capture the nuances of the language."},{"p":1.0,"quote":{"text":"We adopt the Transformer encoder-decoder model proposed by [ref] , the ViT5 [ref] checkpoints, and T5 framework [https://github.com/google-research/text-to-text-transfer-transformer] implemented by [ref] .","citing":"2210.05598","section":"Model Architecture"},"citing":[{"self":false,"year":2022,"title":"Enriching Biomedical Knowledge for Low-resource Language Through Large-Scale Translation","arxiv_id":"2210.05598"}],"verified":true,"paraphrase":"Adopted the Transformer encoder-decoder model proposed by this paper."},{"p":1.0,"quote":{"text":"In our framework, we utilize the following methods for molecular graphs: GradInput [ref] , GradCAM [ref] , Integrated Gradients [ref] , CAM [ref] , MCTS [ref] and Attention weights [ref] .","citing":"2107.04119","section":"XAI methods"},"citing":[{"self":false,"year":2021,"title":"Quantitative Evaluation of Explainable Graph Neural Networks for Molecular Property Prediction","arxiv_id":"2107.04119"}],"verified":true,"paraphrase":"Utilized attention weights among its methods for molecular graphs."},{"p":1.0,"quote":{"text":"We utilize a multi-headed self-attention [ref] mechanism to create an embedding of the state of the vehicle that captures the influence that the surrounding vehicles, as well as static objects, have on the state of the vehicle.","citing":"2506.08957","section":"Attention-Based Trajectory Prediction for Signalized Intersection"},"citing":[{"self":false,"year":2025,"title":"IntTrajSim: Trajectory Prediction for Simulating Multi-Vehicle driving at Signalized Intersections","arxiv_id":"2506.08957"}],"verified":true,"paraphrase":"Utilized a multi-headed self-attention mechanism to create an embedding of the vehicle's state."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2020,"title":"Experiments with LVT and FRE for Transformer model","arxiv_id":"2004.12495"}],"verified":true,"paraphrase":"Used the Transformer model with the base setting in its experiments."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2023,"title":"Textual Augmentation Techniques Applied to Low Resource Machine Translation: Case of Swahili","arxiv_id":"2306.07414"}],"verified":true,"paraphrase":"Trained all its models using this paper's transformer architecture."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2021,"title":"Translating synthetic natural language to database queries: a polyglot deep learning framework","arxiv_id":"2104.07010"}],"verified":true,"paraphrase":"Utilized the Transformer architecture to train a model to predict attributes, classes and constraints from a user query."},{"p":1.0,"quote":null,"citing":[{"self":false,"year":2024,"title":"Towards more efficient agricultural practices via transformer-based crop type classification","arxiv_id":"2411.02627"}],"verified":true,"paraphrase":"Used a three-layer transformer model based on upstream code."}],"papers_all":1189,"self_papers":0,"self_entries":[]}],"counts":{"by_kind":{"extends":17,"contests":2,"data_used":9,"comparison":55,"method_or_tool":1189,"supports_this_paper":25},"dropped":{"relied_on":0,"unknown_ids":0,"account_lines":0,"contest_vetoed":0,"below_threshold":0,"contest_drawback":0,"contest_without_basis":0,"contest_not_one_sentence":0,"contest_on_group_citation":0},"papers_read":40,"self_papers":12,"citing_papers":4179,"sentences_read":40,"citing_sentences":6882,"unambiguous_read":40,"typed_independent":1253,"independent_papers":4167,"by_kind_independent":{"extends":16,"contests":2,"data_used":8,"comparison":53,"method_or_tool":1181,"supports_this_paper":25}},"input_sha256":"b6288429834cab8a972d3a4052afae9834f0fd417378af65ff1bac9850700a64","method":"For a paper with five or more citing arXiv papers, Pith selected up to 40 citing sentences its citation classifier typed, read those whose citation markers point at this paper alone, and wrote one paraphrase for each later paper in each group, each resting on one sentence. A later paper is listed under a label only when the classifier gives one of its sentences that label at intent probability 0.8 or more (0.95 for support), Pith's reading of the sentence finds that the citing paper itself did it, and a second check of the paraphrase against the sentence passes. The section lists uses only: building on the paper, using its method, tool, data or values, comparing with it and citing it as support. It does not list criticism or disagreement, so their absence says nothing about how the paper has been received. Code counted the papers, wrote the summary from the listed entries, set the evidence level and linked each entry to its citing paper. Quotes are copied from the citing sentence and shown only when the citing paper's licence is CC BY, CC BY-SA or CC0; otherwise the paraphrase is shown. Papers that share an author with this one are listed apart. A paper with fewer citing papers gets a line counted by code, which names uses only: a later paper whose only label is a contest is counted as not described.","line":null,"summary":"13 later papers by other authors are listed below: 13 used its method or tool."},"_links":{"html":"https://pith.science/paper/1706.03762"}}