- 综合排序
- 最热优先
- 最新优先
- 来自专栏饶文津的专栏
【Codeforces 707C】Pythagorean Triples（找规律）
一边长为a的直角三角形，a^2=c^2-b^2。可以发现1、4、9、16、25依次差3、5、7、9...，所以任何一条长度为奇数的边a，a^2还是奇数，那么c=a^2/2,b=c+1。我们还可以发现，1、4、9、16、25、36各项差为8、12、16、20，偶数的平方是4的倍数，那么c=a^2/4-1,b=a^2/4+1。
48850发布于 2020-06-02 - 来自专栏活动
基于ComplEx模型的知识图谱嵌入详解
# 假设我们已经下载并清洗了DBpedia数据集 dbpedia_data = pd.read_csv('dbpedia_triples.csv') dbpedia_triples = dbpedia_data dbpedia_triples['subject']).union(set(dbpedia_triples['object'])) relation_set = set(dbpedia_triples[ ((head_index, relation_index, tail_index)) # 将数据拆分为训练集和测试集 train_triples, test_triples = train_test_split ): self.triples = triples def __len__(self): return len(self.triples) def _ ): self.triples = triples def __len__(self): return len(self.triples) def _
91210编辑于 2024-09-21 - 来自专栏机器之心
探索「老药新用」最短路径：亚马逊AI Lab开源大规模药物重定位知识图谱DRKG
() seed = np.arange(num_triples) np.random.shuffle(seed) train_cnt = int(num_triples * 0.9) valid_cnt = int(num_triples * 0.05) train_set = seed[:train_cnt] train_set = train_set.tolist() valid_set = seed [idx][0], triples[idx][1], triples[idx][2])) with open("train/drkg_valid.tsv", 'w+') as f: for idx in valid_set: f.writelines("{}\t{}\t{}\n".format(triples[idx][0], triples[idx][1], triples[idx][ [idx][0], triples[idx][1], triples[idx][2])) 随后直接调用 DGL-KE 软件包的命令行进行 DRKG 知识图谱的低纬嵌入向量表示训练，案例中我们选用 TransE_l2
1.2K21发布于 2020-06-16 - 来自专栏WebJ2EE
【RDF】：RDF、Turtle、rdflib
Triples 2.7. IRIs 2.8. Blank Nodes 2.9. Fragment Identifiers 3. RDF 1.1 Turtle 语法 3.1. Turtle 是一种 RDF 图的序列化方式，它比 RDF/XML 序列化方式紧凑，比 N-Triples 可读性好。 2. RDF 1.1 语法 2.1. A set of such triples is called an RDF graph. RDF Graphs An RDF graph is a set of RDF triples. 2.6. An RDF graph is made up of triples consisting of a subject, predicate and object.
1.9K40发布于 2021-09-02 100天跟着CP学PostgreSQL+AI,第30天 :临床决策支持：知识图谱如何辅助医生诊断？
ON medical_triples USING gin (subject, predicate, object); 数据导入与预处理： python 运行 import psycopg2 = list(g) for i in tqdm(range(0, len(triples), batch_size)): batch = triples[i:i+batch_size] insert_query = """ INSERT INTO medical_triples (subject, predicate, object) VALUES (%s, %s, ://example.org/relations#name' LEFT JOIN medical_triples e ON d.subject = e.subject AND e.predicate = d JOIN medical_triples ds ON d.subject = ds.subject AND ds.predicate = 'http://example.org/relations
44510编辑于 2025-08-27- 来自专栏活动
『1024 | 码项目』知识图谱嵌入与知识迁移的结合指南
将三元组转换为DataFrame Args: triples: 三元组列表 Returns: = creator.create_triples(is_source=True) source_df = creator.create_dataframe(source_triples) # 创建目标知识图谱数据 target_triples = creator.create_triples(is_source=False) target_df = creator.create_dataframe(target_triples) # 验证数据 if not creator.validate_data ([head, relation, tail]) source_tensor = torch.tensor(source_triples, dtype
72820编辑于 2024-11-04 - 来自专栏编程小白狼
勾股定理：从几何直观到高效计算的数学基石
六、勾股定理在算法中的应用 6.1 寻找勾股数 勾股数是满足a²+b²=c²的正整数三元组(a,b,c)： def find_pythagorean_triples(limit): """ 找出所有边长不超过limit的勾股数 参数: limit: 最大边长限制 返回: 勾股数列表 """ triples ((a, b, c)) return triples # 找出100以内的勾股数 triples = find_pythagorean_triples(100) print(f"100以内共有 {len(triples)} 组勾股数") print("前10组:", triples[:10]) 6.2 在密码学中的应用 某些加密算法利用勾股数的性质，例如基于勾股数的公钥密码系统： def generate_pythagorean_key ((k*a, k*b, k*c)) k += 1 return sorted(set(triples)) # 使用示例 toolkit
62010编辑于 2025-12-18 - 来自专栏活动
『1024 | 码项目』在低资源环境下训练知识图谱嵌入的详细指南
import torchimport torch.nn as nnimport numpy as np# 模拟一个简单的知识图谱数据集，包括头实体、关系和尾实体triples = [ ("Alice # 定义负采样函数def negative_sample(triples, entity_count): neg_triples = [] for h, r, t in triples: import random# 示例三元组 (药物A, 治疗, 疾病B)triples = [("药物A", "治疗", "疾病B")]# 生成反向三元组augmented_triples = []for h, r, t in triples: augmented_triples.append((t, "被治疗", h))# 添加噪声数据noise_triple = ("药物C", "治疗", " 疾病D")if random.random() > 0.5: augmented_triples.append(noise_triple)print(augmented_triples)2 参数共享与迁移学习在数据稀缺的情况下
48520编辑于 2024-10-23 - 来自专栏DeepHub IMBA
整合文本和知识图谱嵌入提升RAG的性能
with StanfordOpenIE() as client: triples = client.annotate(text) for triple in triples: print(triple) cleaned_triples = [(subject.lower(), relation.lower(), object.lower()) for (subject , relation, object) in triples] print("Cleaned Triples:", cleaned_triples) 上述代码的输出为 ('Hawaii', # Create a directed graph knowledge_graph = nx.DiGraph() # Add nodes and edges from cleaned triples for (subject, relation, object) in cleaned_triples: knowledge_graph.add_edge(subject, object,
93010编辑于 2024-05-10 - 来自专栏一道题做一宿
你有被三数之和难倒吗
high = nums.length - 1; while (low < high) { // 判断是否满足条件 triples = nums[i] + nums[low] + nums[high]; if (triples == requiredSum) { return true; } // 如果小于requiredSum，把低位指针往后移，以让triples更大些 else if (triples < requiredSum) { low++; } // 把高位指针往前移 ，以让triples更小些 else { high--; } }
51020编辑于 2022-04-11 - 来自专栏WebJ2EE
【RDF】：Reification（RDF的边上能搞属性吗？）
intended meaning of this vocabulary is to allow an RDF graph to act as metadata describing other RDF triples Since the relation between triples and reifications of triples in any RDF graph or graphs need not be
1.1K20发布于 2021-09-26 - 来自专栏ceshiren0001
用LightRAG打造智能问答系统：知识图谱与RAG的融合应用实践
= kg_loader.load_from_json("data/kg_data.json")# 或者从CSV文件创建knowledge_graph = kg_loader.create_from_triples ( "data/triples.csv", entity_col="head", relation_col="relation", object_col="tail")3.2 构建知识图谱索引 1000, expire_after=3600 # 1小时缓存)五、实际应用案例5.1 学术文献问答系统# 构建学术知识图谱academic_kg = kg_loader.load_from_triples ( "academic_triples.csv", entity_col="concept", relation_col="relationship", object_col="
1K00编辑于 2025-08-27 - 来自专栏信数据得永生
NumPy Cookbook 带注释源码 六、NumPy 特殊数组与通用函数
start, today) return np.array([q[4] for q in quotes]) # 获取 AAPL 一年的收盘价 close = get_close('AAPL') triples = np.arange(0, len(close), 3) print("Triples", triples[:10], "...") # Triples [ 0 3 6 9 12 15 18 # sign 中每隔三个元素变为 -1 signs[triples] = -1 print("Signs", signs[:10], "...") # Signs [-1. 1. 1. -1
76530发布于 2019-02-15 - 来自专栏CSDN旧文
Codeforce 1255 Round #601 (Div. 2) C. League of Leesins （大模拟）
Formally, Bob created an array qq of n−2n−2 triples, where qi=(pi,pi+1,pi+2)qi=(pi,pi+1,pi+2) for each Bob can then rearrange the numbers within each triple and the positions of the triples to get [(4,3,2 3,3,5)] is not a valid rearrangement of qq, as Bob is not allowed to swap numbers belong to different triples Remember, that the numbers within each triple can be rearranged and also the positions of the triples
29920发布于 2020-10-28 - 来自专栏机器学习入门
Pyramid Transition Matrix
We also start with a list of allowed triples allowed. X Y Z This works because (‘X’, ‘Y’, ‘D’), (‘Y’, ‘Z’, ‘E’), and (‘D’, ‘E’, ‘A’) are allowed triples Note that there could be allowed triples (A, B, C) and (A, B, D) with C != D.
82230发布于 2019-05-26 从非结构化文本到交互式知识图谱 —— 利用大语言模型（LLMs）
(500 words) Processing chunk 2/3(500 words) Processing chunk 3/3(66 words) Extracted a total of 73 triples = PHASE 2: ENTITY STANDARDIZATION ================================================== Startingwith73 triples and106 unique entities Standardizing entity names across all triples... entities into101 standard forms After standardization:65 triples and72 unique entities ============ Saved raw knowledge graph data to mydocument.json Processing116 triples for visualization Found72 unique
1800编辑于 2026-03-13- 来自专栏活动
基于RotatE模型的知识图谱嵌入技术
MovieB': 1, 'DirectorC': 2, 'ActorD': 3}relations = {'likes': 0, 'directed_by': 1, 'acted_in': 2}# 三元组数据triples MovieB'), ('MovieB', 'directed_by', 'DirectorC'), ('DirectorC', 'acted_in', 'ActorD')]# 将三元组转换为索引triples_idx = [(entities[h], relations[r], entities[t]) for h, r, t in triples]RotatE模型实现import torchimport torch.nn torch.mean(torch.clamp(predictions - labels, min=0))# 创建训练数据heads = torch.tensor([h for h, r, t in triples_idx ])relations = torch.tensor([r for h, r, t in triples_idx])tails = torch.tensor([t for h, r, t in triples_idx
96331编辑于 2024-09-28 - 来自专栏信数据得永生
NumPy 秘籍中文第二版：六、特殊数组和通用函数
操作步骤 我们将使用包含负数的数组的对数： 创建一个数组，该数组包含可被三除的数字： triples = np.arange(0, len(close), 3) print("Triples", triples signs[triples] = -1 print("Signs", signs[:10], "...") 这应该为AAPL打印以下输出： Triples [ 0 3 6 9 12 15 18 21 24 27] ... = np.arange(0, len(close), 3) print("Triples", triples[:10], "...") signs[triples] = -1 print("Signs", signs[:10], "...")
82610编辑于 2023-04-17 - 来自专栏活动
知识图谱的基础构建指南
\n"# 解析并清洗数据，转换为三元组格式triples = []for line in raw_data: match = re.match(pattern, line) if match : book, author, publisher = match.groups() triples.append((book, "作者", author)) triples.append((book, "出版社", publisher))# 打印清洗后的三元组数据for triple in triples: print(triple)我们使用正则表达式 连接到 Neo4j 数据库graph = Graph("bolt://localhost:7687", auth=("neo4j", "password"))# 创建节点和关系for triple in triples
2.2K20编辑于 2024-09-13 【杭电oj】1334 - Perfect Cubes（水）
There do exist several values of a which can be produced from multiple distinct sets of b, c, and d triples In these cases, the triples with the smaller b values should be listed first.
27110编辑于 2025-08-26
腾讯云开发者
Copyright © 2013 - 2026 Tencent Cloud. All Rights Reserved. 腾讯云 版权所有
深圳市腾讯计算机系统有限公司 ICP备案/许可证号：粤B2-20090059 
粤公网安备44030502008569号
腾讯云计算（北京）有限责任公司 京ICP证150476号 | 京ICP备11018762号