File size: 12,844 Bytes
448d6a5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
"""Build a self-contained reproduction poster (poster.html -> poster.png data-URI)
and poster_embed.html with accessible hotspots linking to the logbook claim pages.
Figures are rendered with matplotlib from the result JSONs (no external assets)."""
import os, sys, json, base64, io
import numpy as np
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt

BASE = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
OUT = os.path.join(BASE, "outputs", "poster")
os.makedirs(OUT, exist_ok=True)

GEN = "Qwen/Qwen2.5-32B-Instruct-AWQ"
ACCENT = "#2D5F8B"; DEEP = "#1F4566"; EMPH = "#C9A24A"; GREEN = "#2ca02c"; RED = "#d62728"
SLUGS = {
    "c1": "claim-1-mindflow-formulates-research-ideation-as-graph-structured-flow-composed-of-modular-thinking-operators-and-probabilistic-mind-supernet",
    "c2": "claim-2-mindflow-shows-superiority-as-explicit-controllable-and-optimizable-research-idea-innovator-across-diverse-topics",
    "c3": "claim-3-tournament-based-relative-ranking-enables-the-controller-to-progressively-favor-higher-quality-thinking-flows",
}

def load(claim):
    p = os.path.join(BASE, "outputs", claim, f"{claim}_results.json")
    return json.load(open(p)) if os.path.exists(p) else None

def datauri(fig):
    b = io.BytesIO(); fig.savefig(b, format="png", dpi=150, bbox_inches="tight", facecolor="white")
    plt.close(fig); b.seek(0)
    return "data:image/png;base64," + base64.b64encode(b.read()).decode()

def fig_heatmap(c1):
    ops = c1["ctrl_ops"]; heat = c1["heatmap"]; doms = list(heat.keys())
    Z = np.array([[heat[d][o] for o in ops] for d in doms])
    fig, ax = plt.subplots(figsize=(5.2, 3.4))
    im = ax.imshow(Z, cmap="viridis", vmin=0, vmax=1, aspect="auto")
    ax.set_xticks(range(len(ops))); ax.set_xticklabels(ops, rotation=40, ha="right", fontsize=8)
    ax.set_yticks(range(len(doms))); ax.set_yticklabels(doms, fontsize=8)
    fig.colorbar(im, ax=ax, fraction=0.035).ax.tick_params(labelsize=7)
    ax.set_title("Mind-supernet P(include operator) by domain", fontsize=10, color=DEEP, weight="bold")
    return datauri(fig)

def fig_claim2(c2):
    methods = [k for k in c2 if not k.startswith("_")]
    order = sorted(methods, key=lambda m: c2[m]["agg"]["Overall"])
    names = order; ov = [c2[m]["agg"]["Overall"] for m in order]
    cols = [EMPH if m == "MindFlow" else ACCENT for m in order]
    fig, ax = plt.subplots(figsize=(5.2, 3.4))
    ax.barh(names, ov, color=cols)
    for i, v in enumerate(ov): ax.text(v + 0.005, i, f"{v:.3f}", va="center", fontsize=8)
    ax.set_xlabel("Overall win-rate MOScore vs expert", fontsize=9)
    ax.set_title("Claim 2 — MindFlow best aggregate", fontsize=10, color=DEEP, weight="bold")
    ax.tick_params(labelsize=8); ax.set_xlim(0, max(ov) * 1.18)
    return datauri(fig)

def fig_claim3(c3):
    fig, ax = plt.subplots(figsize=(5.2, 3.4))
    for mode, col in [("tournament", GREEN), ("pointwise", RED)]:
        if mode in c3:
            h = c3[mode]["hist"]
            ax.plot(h["eval_iter"], h["eval_overall"], "-o", color=col, label=mode, lw=2, ms=4)
    ax.set_xlabel("Optimization iteration", fontsize=9); ax.set_ylabel("Held-out MOScore", fontsize=9)
    ax.set_title("Claim 3 — tournament vs pointwise", fontsize=10, color=DEEP, weight="bold")
    ax.legend(fontsize=8); ax.tick_params(labelsize=8); ax.grid(alpha=0.3)
    return datauri(fig)

def fig_reward(c3):
    fig, ax = plt.subplots(figsize=(5.2, 2.7))
    for mode, col in [("tournament", GREEN), ("pointwise", RED)]:
        if mode in c3:
            rs = c3[mode]["hist"]["reward_std"]
            ax.plot(range(1, len(rs) + 1), rs, "-", color=col, label=f"{mode} (μ={np.mean(rs):.2f})", lw=2)
    ax.set_xlabel("Iteration", fontsize=9); ax.set_ylabel("Reward std", fontsize=9)
    ax.set_title("Reward discrimination (↑ = less judgment collapse)", fontsize=9, color=DEEP, weight="bold")
    ax.legend(fontsize=8); ax.tick_params(labelsize=8); ax.grid(alpha=0.3)
    return datauri(fig)

def build():
    c1, c2, c3 = load("claim1"), load("claim2"), load("claim3")
    mf = c2["MindFlow"]["agg"]["Overall"]
    methods = [k for k in c2 if not k.startswith("_")]
    bb_name = max((m for m in methods if m != "MindFlow"), key=lambda m: c2[m]["agg"]["Overall"])
    bb = c2[bb_name]["agg"]["Overall"]
    t_ov = c3["tournament"]["hist"]["eval_overall"]; p_ov = c3["pointwise"]["hist"]["eval_overall"]
    t_rstd = float(np.mean(c3["tournament"]["hist"]["reward_std"]))
    p_rstd = float(np.mean(c3["pointwise"]["hist"]["reward_std"]))
    imgs = {"heat": fig_heatmap(c1), "c2": fig_claim2(c2), "c3": fig_claim3(c3), "rew": fig_reward(c3)}

    html = f"""<!-- poster_embed.html source poster -->
<div id="poster" style="width:1600px;height:900px;box-sizing:border-box;font-family:-apple-system,Segoe UI,Helvetica,Arial,sans-serif;background:#fff;color:#1a2733;padding:26px 30px;position:relative;border:1px solid #e5e9ee;">
  <div style="border-bottom:6px solid {ACCENT};padding-bottom:10px;margin-bottom:14px;">
    <div style="font-size:31px;font-weight:800;color:{DEEP};line-height:1.1;">MindFlow: Mind Supernet Powered Thinking Flows for Research Idea Innovation</div>
    <div style="font-size:15px;color:#4a5a6a;margin-top:5px;">Independent reproduction of ICML 2026 #1894 (OpenReview GgINST3Qgc) · open backbone <b>{GEN}</b> via vLLM · all three core claims reproduce at mechanism scale</div>
  </div>
  <div style="display:grid;grid-template-columns:1.02fr 1fr 1fr;gap:16px;height:720px;">
    <!-- Col 1: method -->
    <div style="display:flex;flex-direction:column;gap:12px;">
      <div style="background:#F4F8FB;border:1px solid #dde6ee;border-radius:10px;padding:12px 14px;">
        <div style="font-size:17px;font-weight:700;color:{DEEP};margin-bottom:6px;">The idea</div>
        <div style="font-size:13.5px;line-height:1.45;">Research ideation is <b>open-ended</b> and <b>multi-objective</b> (novelty · plausibility · feasibility). MindFlow makes the thinking process <b>explicit, controllable & optimizable</b>:</div>
        <ul style="font-size:13px;line-height:1.5;margin:8px 0 0 16px;padding:0;">
          <li><b>Thinking flow</b> — a DAG over 8 modular operators (Generate, Divergent, Convergent, Critical, Analogical, Counterfactual, Constraint-Driven, Exit).</li>
          <li><b>Mind supernet</b> — layer-wise operator inclusion probs π<sub>ℓ</sub>(O|topic); a controller samples topic-specific flows.</li>
          <li><b>Tournament ranking</b> — REINFORCE on relative ranks of K flows, not noisy absolute scores.</li>
        </ul>
      </div>
      <div style="background:{DEEP};color:#fff;border-radius:10px;padding:12px 14px;">
        <div style="font-size:16px;font-weight:700;margin-bottom:6px;">Reproduction outcome</div>
        <div style="font-size:13.5px;line-height:1.5;">✔ <b>Claim 1</b> — supernet instantiates a topic-varying distribution over composable flows.<br>
        ✔ <b>Claim 2</b> — trained controller wins aggregate MOScore <b>{mf:.3f}</b> vs best baseline {bb_name} {bb:.3f}.<br>
        ✔ <b>Claim 3</b> — tournament reward (std {t_rstd:.2f}) beats collapsed pointwise (std {p_rstd:.2f}); held-out MOScore {t_ov[0]:.3f}{t_ov[-1]:.3f}.</div>
      </div>
      <div style="background:#FFF7E6;border:1px solid {EMPH};border-radius:10px;padding:10px 14px;font-size:12px;line-height:1.45;">
        <b>Scope & cost.</b> 8-query IdeaBench proxy (vs paper's 3,495 papers); 1× RTX 6000 Ada, ~1 hr, ~\\$1–3. Backbone substitution (open 32B) for the paper's unstated closed LLM; 3-judge panel emulated by one model + order randomization.</div>
    </div>
    <!-- Col 2: claim 1 + claim 2 -->
    <div style="display:flex;flex-direction:column;gap:12px;">
      <div data-target="{SLUGS['c1']}" style="background:#fff;border:1px solid #dde6ee;border-radius:10px;padding:10px 12px;position:relative;">
        <div style="font-size:15px;font-weight:700;color:{DEEP};">Claim 1 · graph-structured supernet <span style="float:right;font-size:11px;color:{ACCENT};">Open details ↗</span></div>
        <img src="{imgs['heat']}" style="width:100%;margin-top:6px;border-radius:6px;"/>
        <div style="font-size:12px;color:#456;margin-top:4px;">Non-uniform, topic-varying operator preferences → a genuine probabilistic supernet.</div>
      </div>
      <div data-target="{SLUGS['c2']}" style="background:#fff;border:1px solid #dde6ee;border-radius:10px;padding:10px 12px;position:relative;">
        <div style="font-size:15px;font-weight:700;color:{DEEP};">Claim 2 · superiority across topics <span style="float:right;font-size:11px;color:{ACCENT};">Open details ↗</span></div>
        <img src="{imgs['c2']}" style="width:100%;margin-top:6px;border-radius:6px;"/>
      </div>
    </div>
    <!-- Col 3: claim 3 -->
    <div data-target="{SLUGS['c3']}" style="display:flex;flex-direction:column;gap:10px;background:#fff;border:1px solid #dde6ee;border-radius:10px;padding:10px 12px;position:relative;">
      <div style="font-size:15px;font-weight:700;color:{DEEP};">Claim 3 · tournament ranking optimizes the controller <span style="float:right;font-size:11px;color:{ACCENT};">Open details ↗</span></div>
      <img src="{imgs['c3']}" style="width:100%;border-radius:6px;"/>
      <img src="{imgs['rew']}" style="width:100%;border-radius:6px;"/>
      <div style="font-size:12px;color:#456;">Relative ranking always spreads candidates across ranks 0..K-1 → stable gradient; absolute scoring collapses into a narrow band ('judgment collapse') and barely moves the controller.</div>
    </div>
  </div>
  <div style="position:absolute;bottom:14px;left:30px;right:30px;border-top:2px solid #e5e9ee;padding-top:8px;font-size:11.5px;color:#5a6a7a;display:flex;justify-content:space-between;">
    <span>Reproduction logbook · Trackio · backbone {GEN} · encoder all-MiniLM-L6-v2</span>
    <span>ICML 2026 open-reproduction challenge</span>
  </div>
</div>"""
    open(os.path.join(OUT, "poster.html"), "w").write(html)
    print("wrote poster.html", flush=True)
    return html

def render_png():
    from playwright.sync_api import sync_playwright
    src = os.path.join(OUT, "poster.html")
    png = os.path.join(OUT, "poster.png")
    full = "<!doctype html><html><head><meta charset='utf-8'></head><body style='margin:0'>" + open(src).read() + "</body></html>"
    tmp = os.path.join(OUT, "_poster_full.html"); open(tmp, "w").write(full)
    with sync_playwright() as p:
        b = p.chromium.launch(); pg = b.new_page(viewport={"width": 1600, "height": 900}, device_scale_factor=2)
        pg.goto("file://" + tmp); pg.wait_for_timeout(400)
        el = pg.query_selector("#poster"); el.screenshot(path=png)
        b.close()
    print("wrote poster.png", os.path.getsize(png), flush=True)
    return png

def build_embed():
    png = os.path.join(OUT, "poster.png")
    uri = "data:image/png;base64," + base64.b64encode(open(png, "rb").read()).decode()
    # hotspots (percentages of the 1600x900 canvas) over the three claim regions
    spots = [
        {"slug": SLUGS["c1"], "label": "Claim 1 details", "x": 35.0, "y": 20.0, "w": 31.5, "h": 33.0},
        {"slug": SLUGS["c2"], "label": "Claim 2 details", "x": 35.0, "y": 55.0, "w": 31.5, "h": 33.0},
        {"slug": SLUGS["c3"], "label": "Claim 3 details", "x": 67.5, "y": 20.0, "w": 31.5, "h": 68.0},
    ]
    btns = "\n".join(
        f'<a class="hot" href="#/{s["slug"]}" target="_top" aria-label="{s["label"]}" '
        f'style="left:{s["x"]}%;top:{s["y"]}%;width:{s["w"]}%;height:{s["h"]}%;" title="{s["label"]}">'
        f'<span class="pill">Open details ↗</span></a>' for s in spots)
    html = f"""<!-- poster_embed.html : self-contained reproduction poster with accessible hotspots -->
<div class="poster-embed" style="position:relative;max-width:100%;margin:0 auto;">
  <style>
    .poster-embed img{{width:100%;display:block;border-radius:8px;}}
    .poster-embed .hot{{position:absolute;display:flex;align-items:flex-start;justify-content:flex-end;
      border:2px solid transparent;border-radius:8px;text-decoration:none;transition:.15s;}}
    .poster-embed .hot:hover,.poster-embed .hot:focus{{border-color:#2D5F8B;background:rgba(45,95,139,.08);outline:none;}}
    .poster-embed .pill{{margin:6px;background:#2D5F8B;color:#fff;font:600 11px/1 -apple-system,Segoe UI,Arial;
      padding:4px 8px;border-radius:12px;opacity:.85;}}
    .poster-embed .hot:hover .pill,.poster-embed .hot:focus .pill{{opacity:1;}}
  </style>
  <img src="{uri}" alt="MindFlow reproduction poster"/>
  {btns}
</div>"""
    open(os.path.join(OUT, "poster_embed.html"), "w").write(html)
    print("wrote poster_embed.html", os.path.getsize(os.path.join(OUT, "poster_embed.html")), flush=True)

if __name__ == "__main__":
    build(); render_png(); build_embed()
    print("poster done ->", OUT)