File size: 4,745 Bytes
2c278e8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b0eaf65
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8" />
  <meta name="viewport" content="width=device-width, initial-scale=1.0" />
  <title>Medical Research Architecture</title>
  <style>
    body {
      font-family: Arial, sans-serif;
      background: #f7f9fa;
      margin: 0;
      padding: 40px;
    }
    .step {
      background: #ffffff;
      border-left: 6px solid #4CAF50;
      padding: 20px;
      margin-bottom: 30px;
      border-radius: 10px;
      box-shadow: 0 2px 6px rgba(0, 0, 0, 0.1);
    }
    .step h2 {
      margin-top: 0;
    }
    .agents {
      display: flex;
      flex-wrap: wrap;
      gap: 20px;
      margin-top: 15px;
    }
    .agent-box {
      background: #e9f5ff;
      border: 1px solid #cce7ff;
      border-radius: 8px;
      padding: 10px 15px;
      flex: 1 1 30%;
      min-width: 250px;
      box-shadow: 0 1px 4px rgba(0, 0, 0, 0.05);
    }
    .arrow {
      text-align: center;
      margin: 10px 0;
      font-size: 20px;
    }
    .note {
      font-style: italic;
      font-size: 0.9em;
      color: #555;
      margin-top: 10px;
    }
  </style>
</head>
<body>
  <div class="step">
    <h2>STEP 1: Parallelized Data Collection</h2>
    <p><strong>Supervisor Agent</strong> dispatches specialized agents to fetch relevant data in parallel.</p>
    <div class="agents">
      <div class="agent-box">
        <strong>GBQ Research Agent</strong><br/>
        - Fetches structured internal datasets from BigQuery.
      </div>
      <div class="agent-box">
        <strong>DailyMed Agent</strong><br/>
        - Retrieves verified drug information and monographs.
      </div>
      <div class="agent-box">
        <strong>Web Research Agent</strong><br/>
        - Crawls and scrapes open medical research and news.
      </div>
    </div>
  </div>

  <div class="step">
    <h2>STEP 2: Data Synthesis & Gap Analysis</h2>
    <p>The <strong>Synthesizer Agent</strong> merges the raw findings into a preliminary brief, resolving conflicts and identifying gaps.</p>
    <div class="agents">
      <div class="agent-box">
        <strong>Input</strong><br/>
        - Data from all Step 1 Agents.
      </div>
      <div class="arrow"></div>
      <div class="agent-box">
        <strong>Synthesizer Agent</strong><br/>
        - Creates: <em>Preliminary Research Brief</em>
        <div class="note">Highlights inconsistencies and missing data points.</div>
      </div>
    </div>
  </div>

  <div class="step">
    <h2>STEP 3: Collaborative Summarization & Voting</h2>
    <p>Multiple LLM agents draft independent summaries, then vote to reach a unified consensus.</p>
    <div class="agents">
      <div class="agent-box">
        <strong>Summarizer Agent A</strong><br/>
        - Powered by Claude<br/>
        - Drafts version 1
      </div>
      <div class="agent-box">
        <strong>Summarizer Agent B</strong><br/>
        - Powered by Gemini<br/>
        - Drafts version 2
      </div>
      <div class="agent-box">
        <strong>Summarizer Agent C</strong><br/>
        - Powered by GPT-4o<br/>
        - Drafts version 3
      </div>
    </div>
    <div class="arrow"></div>
    <div class="agents">
      <div class="agent-box">
        <strong>Critique & Voting Agent</strong><br/>
        - Compares drafts against a rubric<br/>
        - Creates: <em>Consensus Summary</em>
        <div class="note">Integrates strengths from all drafts.</div>
      </div>
    </div>
  </div>

  <div class="step">
    <h2>STEP 4: Human-in-the-Loop Review</h2>
    <p>Medical researchers review AI-generated content, provide feedback, and request clarification.</p>
    <div class="agents">
      <div class="agent-box">
        <strong>Interface Viewer</strong><br/>
        - Displays Consensus Summary and source info<br/>
        - Enables targeted review by human experts
      </div>
      <div class="arrow"></div>
      <div class="agent-box">
        <strong>Medical Researcher (MR)</strong><br/>
        - Reviews key chunks<br/>
        - Provides corrections, queries, and follow-up requests
      </div>
    </div>
  </div>

  <div class="step">
    <h2>STEP 5: Final Report Generation</h2>
    <p>The final LLM synthesizes validated data into a formal, structured report.</p>
    <div class="agents">
      <div class="agent-box">
        <strong>Report Generation Agent</strong><br/>
        - Powered by GPT-4o / Gemini 1.5 Pro<br/>
        - Inputs: raw data, consensus summary, MR feedback
      </div>
      <div class="arrow"></div>
      <div class="agent-box">
        <strong>Final Research Document</strong><br/>
        - Fully structured, validated, and ready for submission or publication
      </div>
    </div>
  </div>

</body>
</html>